diff --git a/glm/detail/_features.hpp b/glm/detail/_features.hpp index b0aa3036..4aa4ad5c 100644 --- a/glm/detail/_features.hpp +++ b/glm/detail/_features.hpp @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/_features.hpp -/// @date 2013-02-20 / 2013-02-20 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/detail/_fixes.hpp b/glm/detail/_fixes.hpp index 0405f33a..c957562b 100644 --- a/glm/detail/_fixes.hpp +++ b/glm/detail/_fixes.hpp @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/_fixes.hpp -/// @date 2011-02-21 / 2011-11-22 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #include diff --git a/glm/detail/_noise.hpp b/glm/detail/_noise.hpp index e8875976..89403f4a 100644 --- a/glm/detail/_noise.hpp +++ b/glm/detail/_noise.hpp @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/_noise.hpp -/// @date 2013-12-24 / 2013-12-24 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/detail/_swizzle.hpp b/glm/detail/_swizzle.hpp index 2f23d0e1..b589a2be 100644 --- a/glm/detail/_swizzle.hpp +++ b/glm/detail/_swizzle.hpp @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/_swizzle.hpp -/// @date 2006-04-20 / 2011-02-16 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/detail/_swizzle_func.hpp b/glm/detail/_swizzle_func.hpp index 464109d4..4b37edbb 100644 --- a/glm/detail/_swizzle_func.hpp +++ b/glm/detail/_swizzle_func.hpp @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/_swizzle_func.hpp -/// @date 2011-10-16 / 2011-10-16 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/detail/_vectorize.hpp b/glm/detail/_vectorize.hpp index 53cf5b37..a08ed34d 100644 --- a/glm/detail/_vectorize.hpp +++ b/glm/detail/_vectorize.hpp @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/_vectorize.hpp -/// @date 2011-10-14 / 2011-10-14 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/detail/dummy.cpp b/glm/detail/dummy.cpp index 4da61968..af163db8 100644 --- a/glm/detail/dummy.cpp +++ b/glm/detail/dummy.cpp @@ -1,37 +1,10 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/core/dummy.cpp -/// @date 2011-01-19 / 2011-06-15 -/// @author Christophe Riccio /// /// GLM is a header only library. There is nothing to compile. /// dummy.cpp exist only a wordaround for CMake file. /////////////////////////////////////////////////////////////////////////////////// + /* #define GLM_MESSAGES #include diff --git a/glm/detail/func_common.hpp b/glm/detail/func_common.hpp index 08b69c80..871fed6c 100644 --- a/glm/detail/func_common.hpp +++ b/glm/detail/func_common.hpp @@ -1,33 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/func_common.hpp -/// @date 2008-03-08 / 2010-01-26 -/// @author Christophe Riccio /// /// @see GLSL 4.20.8 specification, section 8.3 Common Functions /// @@ -35,7 +7,6 @@ /// @ingroup core /// /// These all operate component-wise. The description is per component. -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/detail/func_common.inl b/glm/detail/func_common.inl index 82d70b5a..0cebb9c9 100644 --- a/glm/detail/func_common.inl +++ b/glm/detail/func_common.inl @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/func_common.inl -/// @date 2008-08-03 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #include "func_vector_relational.hpp" #include "type_vec2.hpp" diff --git a/glm/detail/func_common_simd.inl b/glm/detail/func_common_simd.inl index 308673d4..3afc88b8 100644 --- a/glm/detail/func_common_simd.inl +++ b/glm/detail/func_common_simd.inl @@ -33,6 +33,110 @@ namespace detail static const __m128 GLM_VAR_USED _ps_2pow23 = _mm_set_ps1(8388608.0f); static const __m128 GLM_VAR_USED _ps_1 = _mm_set_ps1(1.0f); +/* + static const __m128 GLM_VAR_USED zero = _mm_setzero_ps(); + static const __m128 GLM_VAR_USED one = _mm_set_ps1(1.0f); + static const __m128 GLM_VAR_USED minus_one = _mm_set_ps1(-1.0f); + static const __m128 GLM_VAR_USED two = _mm_set_ps1(2.0f); + static const __m128 GLM_VAR_USED three = _mm_set_ps1(3.0f); + static const __m128 GLM_VAR_USED pi = _mm_set_ps1(3.1415926535897932384626433832795f); + static const __m128 GLM_VAR_USED hundred_eighty = _mm_set_ps1(180.f); + static const __m128 GLM_VAR_USED pi_over_hundred_eighty = _mm_set_ps1(0.017453292519943295769236907684886f); + static const __m128 GLM_VAR_USED hundred_eighty_over_pi = _mm_set_ps1(57.295779513082320876798154814105f); + + static const ieee754_QNAN absMask; + static const __m128 GLM_VAR_USED abs4Mask = _mm_set_ps1(absMask.f); + + static const __m128 GLM_VAR_USED _epi32_sign_mask = _mm_castsi128_ps(_mm_set1_epi32(static_cast(0x80000000))); + //static const __m128 GLM_VAR_USED _epi32_inv_sign_mask = _mm_castsi128_ps(_mm_set1_epi32(0x7FFFFFFF)); + //static const __m128 GLM_VAR_USED _epi32_mant_mask = _mm_castsi128_ps(_mm_set1_epi32(0x7F800000)); + //static const __m128 GLM_VAR_USED _epi32_inv_mant_mask = _mm_castsi128_ps(_mm_set1_epi32(0x807FFFFF)); + //static const __m128 GLM_VAR_USED _epi32_min_norm_pos = _mm_castsi128_ps(_mm_set1_epi32(0x00800000)); + static const __m128 GLM_VAR_USED _epi32_0 = _mm_set_ps1(0); + static const __m128 GLM_VAR_USED _epi32_1 = _mm_set_ps1(1); + static const __m128 GLM_VAR_USED _epi32_2 = _mm_set_ps1(2); + static const __m128 GLM_VAR_USED _epi32_3 = _mm_set_ps1(3); + static const __m128 GLM_VAR_USED _epi32_4 = _mm_set_ps1(4); + static const __m128 GLM_VAR_USED _epi32_5 = _mm_set_ps1(5); + static const __m128 GLM_VAR_USED _epi32_6 = _mm_set_ps1(6); + static const __m128 GLM_VAR_USED _epi32_7 = _mm_set_ps1(7); + static const __m128 GLM_VAR_USED _epi32_8 = _mm_set_ps1(8); + static const __m128 GLM_VAR_USED _epi32_9 = _mm_set_ps1(9); + static const __m128 GLM_VAR_USED _epi32_127 = _mm_set_ps1(127); + //static const __m128 GLM_VAR_USED _epi32_ninf = _mm_castsi128_ps(_mm_set1_epi32(0xFF800000)); + //static const __m128 GLM_VAR_USED _epi32_pinf = _mm_castsi128_ps(_mm_set1_epi32(0x7F800000)); + + static const __m128 GLM_VAR_USED _ps_1_3 = _mm_set_ps1(0.33333333333333333333333333333333f); + static const __m128 GLM_VAR_USED _ps_0p5 = _mm_set_ps1(0.5f); + static const __m128 GLM_VAR_USED _ps_1 = _mm_set_ps1(1.0f); + static const __m128 GLM_VAR_USED _ps_m1 = _mm_set_ps1(-1.0f); + static const __m128 GLM_VAR_USED _ps_2 = _mm_set_ps1(2.0f); + static const __m128 GLM_VAR_USED _ps_3 = _mm_set_ps1(3.0f); + static const __m128 GLM_VAR_USED _ps_127 = _mm_set_ps1(127.0f); + static const __m128 GLM_VAR_USED _ps_255 = _mm_set_ps1(255.0f); + static const __m128 GLM_VAR_USED _ps_2pow23 = _mm_set_ps1(8388608.0f); + + static const __m128 GLM_VAR_USED _ps_1_0_0_0 = _mm_set_ps(1.0f, 0.0f, 0.0f, 0.0f); + static const __m128 GLM_VAR_USED _ps_0_1_0_0 = _mm_set_ps(0.0f, 1.0f, 0.0f, 0.0f); + static const __m128 GLM_VAR_USED _ps_0_0_1_0 = _mm_set_ps(0.0f, 0.0f, 1.0f, 0.0f); + static const __m128 GLM_VAR_USED _ps_0_0_0_1 = _mm_set_ps(0.0f, 0.0f, 0.0f, 1.0f); + + static const __m128 GLM_VAR_USED _ps_pi = _mm_set_ps1(3.1415926535897932384626433832795f); + static const __m128 GLM_VAR_USED _ps_pi2 = _mm_set_ps1(6.283185307179586476925286766560f); + static const __m128 GLM_VAR_USED _ps_2_pi = _mm_set_ps1(0.63661977236758134307553505349006f); + static const __m128 GLM_VAR_USED _ps_pi_2 = _mm_set_ps1(1.5707963267948966192313216916398f); + static const __m128 GLM_VAR_USED _ps_4_pi = _mm_set_ps1(1.2732395447351626861510701069801f); + static const __m128 GLM_VAR_USED _ps_pi_4 = _mm_set_ps1(0.78539816339744830961566084581988f); + + static const __m128 GLM_VAR_USED _ps_sincos_p0 = _mm_set_ps1(0.15707963267948963959e1f); + static const __m128 GLM_VAR_USED _ps_sincos_p1 = _mm_set_ps1(-0.64596409750621907082e0f); + static const __m128 GLM_VAR_USED _ps_sincos_p2 = _mm_set_ps1(0.7969262624561800806e-1f); + static const __m128 GLM_VAR_USED _ps_sincos_p3 = _mm_set_ps1(-0.468175413106023168e-2f); + static const __m128 GLM_VAR_USED _ps_tan_p0 = _mm_set_ps1(-1.79565251976484877988e7f); + static const __m128 GLM_VAR_USED _ps_tan_p1 = _mm_set_ps1(1.15351664838587416140e6f); + static const __m128 GLM_VAR_USED _ps_tan_p2 = _mm_set_ps1(-1.30936939181383777646e4f); + static const __m128 GLM_VAR_USED _ps_tan_q0 = _mm_set_ps1(-5.38695755929454629881e7f); + static const __m128 GLM_VAR_USED _ps_tan_q1 = _mm_set_ps1(2.50083801823357915839e7f); + static const __m128 GLM_VAR_USED _ps_tan_q2 = _mm_set_ps1(-1.32089234440210967447e6f); + static const __m128 GLM_VAR_USED _ps_tan_q3 = _mm_set_ps1(1.36812963470692954678e4f); + static const __m128 GLM_VAR_USED _ps_tan_poleval = _mm_set_ps1(3.68935e19f); + static const __m128 GLM_VAR_USED _ps_atan_t0 = _mm_set_ps1(-0.91646118527267623468e-1f); + static const __m128 GLM_VAR_USED _ps_atan_t1 = _mm_set_ps1(-0.13956945682312098640e1f); + static const __m128 GLM_VAR_USED _ps_atan_t2 = _mm_set_ps1(-0.94393926122725531747e2f); + static const __m128 GLM_VAR_USED _ps_atan_t3 = _mm_set_ps1(0.12888383034157279340e2f); + static const __m128 GLM_VAR_USED _ps_atan_s0 = _mm_set_ps1(0.12797564625607904396e1f); + static const __m128 GLM_VAR_USED _ps_atan_s1 = _mm_set_ps1(0.21972168858277355914e1f); + static const __m128 GLM_VAR_USED _ps_atan_s2 = _mm_set_ps1(0.68193064729268275701e1f); + static const __m128 GLM_VAR_USED _ps_atan_s3 = _mm_set_ps1(0.28205206687035841409e2f); + + static const __m128 GLM_VAR_USED _ps_exp_hi = _mm_set_ps1(88.3762626647949f); + static const __m128 GLM_VAR_USED _ps_exp_lo = _mm_set_ps1(-88.3762626647949f); + static const __m128 GLM_VAR_USED _ps_exp_rln2 = _mm_set_ps1(1.4426950408889634073599f); + static const __m128 GLM_VAR_USED _ps_exp_p0 = _mm_set_ps1(1.26177193074810590878e-4f); + static const __m128 GLM_VAR_USED _ps_exp_p1 = _mm_set_ps1(3.02994407707441961300e-2f); + static const __m128 GLM_VAR_USED _ps_exp_q0 = _mm_set_ps1(3.00198505138664455042e-6f); + static const __m128 GLM_VAR_USED _ps_exp_q1 = _mm_set_ps1(2.52448340349684104192e-3f); + static const __m128 GLM_VAR_USED _ps_exp_q2 = _mm_set_ps1(2.27265548208155028766e-1f); + static const __m128 GLM_VAR_USED _ps_exp_q3 = _mm_set_ps1(2.00000000000000000009e0f); + static const __m128 GLM_VAR_USED _ps_exp_c1 = _mm_set_ps1(6.93145751953125e-1f); + static const __m128 GLM_VAR_USED _ps_exp_c2 = _mm_set_ps1(1.42860682030941723212e-6f); + static const __m128 GLM_VAR_USED _ps_exp2_hi = _mm_set_ps1(127.4999961853f); + static const __m128 GLM_VAR_USED _ps_exp2_lo = _mm_set_ps1(-127.4999961853f); + static const __m128 GLM_VAR_USED _ps_exp2_p0 = _mm_set_ps1(2.30933477057345225087e-2f); + static const __m128 GLM_VAR_USED _ps_exp2_p1 = _mm_set_ps1(2.02020656693165307700e1f); + static const __m128 GLM_VAR_USED _ps_exp2_p2 = _mm_set_ps1(1.51390680115615096133e3f); + static const __m128 GLM_VAR_USED _ps_exp2_q0 = _mm_set_ps1(2.33184211722314911771e2f); + static const __m128 GLM_VAR_USED _ps_exp2_q1 = _mm_set_ps1(4.36821166879210612817e3f); + static const __m128 GLM_VAR_USED _ps_log_p0 = _mm_set_ps1(-7.89580278884799154124e-1f); + static const __m128 GLM_VAR_USED _ps_log_p1 = _mm_set_ps1(1.63866645699558079767e1f); + static const __m128 GLM_VAR_USED _ps_log_p2 = _mm_set_ps1(-6.41409952958715622951e1f); + static const __m128 GLM_VAR_USED _ps_log_q0 = _mm_set_ps1(-3.56722798256324312549e1f); + static const __m128 GLM_VAR_USED _ps_log_q1 = _mm_set_ps1(3.12093766372244180303e2f); + static const __m128 GLM_VAR_USED _ps_log_q2 = _mm_set_ps1(-7.69691943550460008604e2f); + static const __m128 GLM_VAR_USED _ps_log_c0 = _mm_set_ps1(0.693147180559945f); + static const __m128 GLM_VAR_USED _ps_log2_c0 = _mm_set_ps1(1.44269504088896340735992f); +*/ + GLM_FUNC_QUALIFIER __m128 abs_ps(__m128 x) { return _mm_and_ps(abs4Mask, x); diff --git a/glm/detail/func_exponential.hpp b/glm/detail/func_exponential.hpp index 9277990b..a83de0bb 100644 --- a/glm/detail/func_exponential.hpp +++ b/glm/detail/func_exponential.hpp @@ -1,33 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/func_exponential.hpp -/// @date 2008-08-08 / 2011-06-14 -/// @author Christophe Riccio /// /// @see GLSL 4.20.8 specification, section 8.2 Exponential Functions /// @@ -35,7 +7,6 @@ /// @ingroup core /// /// These all operate component-wise. The description is per component. -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/detail/func_exponential.inl b/glm/detail/func_exponential.inl index d5583922..20d1141f 100644 --- a/glm/detail/func_exponential.inl +++ b/glm/detail/func_exponential.inl @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/func_exponential.inl -/// @date 2008-08-03 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #include "func_vector_relational.hpp" #include "_vectorize.hpp" diff --git a/glm/detail/func_geometric.hpp b/glm/detail/func_geometric.hpp index cdbcc7b8..2f39bf5b 100644 --- a/glm/detail/func_geometric.hpp +++ b/glm/detail/func_geometric.hpp @@ -1,33 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/func_geometric.hpp -/// @date 2008-08-03 / 2011-06-14 -/// @author Christophe Riccio /// /// @see GLSL 4.20.8 specification, section 8.5 Geometric Functions /// @@ -35,7 +7,6 @@ /// @ingroup core /// /// These operate on vectors as vectors, not component-wise. -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/detail/func_geometric.inl b/glm/detail/func_geometric.inl index 7d168ed0..4b845d84 100644 --- a/glm/detail/func_geometric.inl +++ b/glm/detail/func_geometric.inl @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/func_geometric.inl -/// @date 2008-08-03 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #include "func_exponential.hpp" #include "func_common.hpp" diff --git a/glm/detail/func_integer.hpp b/glm/detail/func_integer.hpp index fa32fa1e..bd195a9c 100644 --- a/glm/detail/func_integer.hpp +++ b/glm/detail/func_integer.hpp @@ -1,33 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/func_integer.hpp -/// @date 2010-03-17 / 2011-06-18 -/// @author Christophe Riccio /// /// @see GLSL 4.20.8 specification, section 8.8 Integer Functions /// @@ -37,7 +9,6 @@ /// These all operate component-wise. The description is per component. /// The notation [a, b] means the set of bits from bit-number a through bit-number /// b, inclusive. The lowest-order bit is bit 0. -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/detail/func_integer.inl b/glm/detail/func_integer.inl index e0d5c739..c0a5e9cb 100644 --- a/glm/detail/func_integer.inl +++ b/glm/detail/func_integer.inl @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/func_integer.inl -/// @date 2010-03-17 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #include "type_vec2.hpp" #include "type_vec3.hpp" @@ -387,3 +358,8 @@ namespace detail return detail::compute_findMSB_vec::call(x); } }//namespace glm + +#if GLM_ARCH != GLM_FORCE_PURE && GLM_HAS_ANONYMOUS_UNION && GLM_NOT_BUGGY_VC32BITS +# include "func_integer_simd.inl" +#endif + diff --git a/glm/detail/intrinsic_integer.inl b/glm/detail/func_integer_simd.inl similarity index 71% rename from glm/detail/intrinsic_integer.inl rename to glm/detail/func_integer_simd.inl index 90426e34..6fab1146 100644 --- a/glm/detail/intrinsic_integer.inl +++ b/glm/detail/func_integer_simd.inl @@ -1,30 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2012 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core -/// @file glm/detail/intrinsic_integer.inl -/// @date 2009-05-08 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// +/// @file glm/detail/func_integer_simd.inl namespace glm{ namespace detail @@ -136,4 +111,4 @@ namespace detail return Reg1; } }//namespace detail -}//namespace glms +}//namespace glm diff --git a/glm/detail/func_matrix.inl b/glm/detail/func_matrix.inl index 42b09c83..7dbd1805 100644 --- a/glm/detail/func_matrix.inl +++ b/glm/detail/func_matrix.inl @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/func_matrix.inl -/// @date 2008-03-08 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #include "../geometric.hpp" #include diff --git a/glm/detail/func_matrix_simd.inl b/glm/detail/func_matrix_simd.inl index 80238949..fe008b2e 100644 --- a/glm/detail/func_matrix_simd.inl +++ b/glm/detail/func_matrix_simd.inl @@ -1,8 +1,1050 @@ +/// @ref core +/// @file glm/detail/func_matrix_simd.inl + #include "type_mat4x4.hpp" #include "func_geometric.hpp" -namespace glm +namespace glm{ +namespace detail { + 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); + + template + GLM_FUNC_QUALIFIER matType sse_comp_mul_ps + ( + __m128 const in1[4], + __m128 const in2[4], + __m128 out[4] + ) + { + out[0] = _mm_mul_ps(in1[0], in2[0]); + out[1] = _mm_mul_ps(in1[1], in2[1]); + out[2] = _mm_mul_ps(in1[2], in2[2]); + out[3] = _mm_mul_ps(in1[3], in2[3]); + } + + GLM_FUNC_QUALIFIER void sse_add_ps(__m128 const in1[4], __m128 const in2[4], __m128 out[4]) + { + { + out[0] = _mm_add_ps(in1[0], in2[0]); + out[1] = _mm_add_ps(in1[1], in2[1]); + out[2] = _mm_add_ps(in1[2], in2[2]); + out[3] = _mm_add_ps(in1[3], in2[3]); + } + } + + GLM_FUNC_QUALIFIER void sse_sub_ps(__m128 const in1[4], __m128 const in2[4], __m128 out[4]) + { + { + out[0] = _mm_sub_ps(in1[0], in2[0]); + out[1] = _mm_sub_ps(in1[1], in2[1]); + out[2] = _mm_sub_ps(in1[2], in2[2]); + out[3] = _mm_sub_ps(in1[3], in2[3]); + } + } + + GLM_FUNC_QUALIFIER __m128 sse_mul_ps(__m128 const m[4], __m128 v) + { + __m128 v0 = _mm_shuffle_ps(v, v, _MM_SHUFFLE(0, 0, 0, 0)); + __m128 v1 = _mm_shuffle_ps(v, v, _MM_SHUFFLE(1, 1, 1, 1)); + __m128 v2 = _mm_shuffle_ps(v, v, _MM_SHUFFLE(2, 2, 2, 2)); + __m128 v3 = _mm_shuffle_ps(v, v, _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 m0 = _mm_mul_ps(m[0], v0); + __m128 m1 = _mm_mul_ps(m[1], v1); + __m128 m2 = _mm_mul_ps(m[2], v2); + __m128 m3 = _mm_mul_ps(m[3], v3); + + __m128 a0 = _mm_add_ps(m0, m1); + __m128 a1 = _mm_add_ps(m2, m3); + __m128 a2 = _mm_add_ps(a0, a1); + + return a2; + } + + GLM_FUNC_QUALIFIER __m128 sse_mul_ps(__m128 v, __m128 const m[4]) + { + __m128 i0 = m[0]; + __m128 i1 = m[1]; + __m128 i2 = m[2]; + __m128 i3 = m[3]; + + __m128 m0 = _mm_mul_ps(v, i0); + __m128 m1 = _mm_mul_ps(v, i1); + __m128 m2 = _mm_mul_ps(v, i2); + __m128 m3 = _mm_mul_ps(v, i3); + + __m128 u0 = _mm_unpacklo_ps(m0, m1); + __m128 u1 = _mm_unpackhi_ps(m0, m1); + __m128 a0 = _mm_add_ps(u0, u1); + + __m128 u2 = _mm_unpacklo_ps(m2, m3); + __m128 u3 = _mm_unpackhi_ps(m2, m3); + __m128 a1 = _mm_add_ps(u2, u3); + + __m128 f0 = _mm_movelh_ps(a0, a1); + __m128 f1 = _mm_movehl_ps(a1, a0); + __m128 f2 = _mm_add_ps(f0, f1); + + return f2; + } + + GLM_FUNC_QUALIFIER void sse_mul_ps(__m128 const in1[4], __m128 const in2[4], __m128 out[4]) + { + { + __m128 e0 = _mm_shuffle_ps(in2[0], in2[0], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 e1 = _mm_shuffle_ps(in2[0], in2[0], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 e2 = _mm_shuffle_ps(in2[0], in2[0], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 e3 = _mm_shuffle_ps(in2[0], in2[0], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 m0 = _mm_mul_ps(in1[0], e0); + __m128 m1 = _mm_mul_ps(in1[1], e1); + __m128 m2 = _mm_mul_ps(in1[2], e2); + __m128 m3 = _mm_mul_ps(in1[3], e3); + + __m128 a0 = _mm_add_ps(m0, m1); + __m128 a1 = _mm_add_ps(m2, m3); + __m128 a2 = _mm_add_ps(a0, a1); + + out[0] = a2; + } + + { + __m128 e0 = _mm_shuffle_ps(in2[1], in2[1], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 e1 = _mm_shuffle_ps(in2[1], in2[1], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 e2 = _mm_shuffle_ps(in2[1], in2[1], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 e3 = _mm_shuffle_ps(in2[1], in2[1], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 m0 = _mm_mul_ps(in1[0], e0); + __m128 m1 = _mm_mul_ps(in1[1], e1); + __m128 m2 = _mm_mul_ps(in1[2], e2); + __m128 m3 = _mm_mul_ps(in1[3], e3); + + __m128 a0 = _mm_add_ps(m0, m1); + __m128 a1 = _mm_add_ps(m2, m3); + __m128 a2 = _mm_add_ps(a0, a1); + + out[1] = a2; + } + + { + __m128 e0 = _mm_shuffle_ps(in2[2], in2[2], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 e1 = _mm_shuffle_ps(in2[2], in2[2], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 e2 = _mm_shuffle_ps(in2[2], in2[2], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 e3 = _mm_shuffle_ps(in2[2], in2[2], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 m0 = _mm_mul_ps(in1[0], e0); + __m128 m1 = _mm_mul_ps(in1[1], e1); + __m128 m2 = _mm_mul_ps(in1[2], e2); + __m128 m3 = _mm_mul_ps(in1[3], e3); + + __m128 a0 = _mm_add_ps(m0, m1); + __m128 a1 = _mm_add_ps(m2, m3); + __m128 a2 = _mm_add_ps(a0, a1); + + out[2] = a2; + } + + { + //(__m128&)_mm_shuffle_epi32(__m128i&)in2[0], _MM_SHUFFLE(3, 3, 3, 3)) + __m128 e0 = _mm_shuffle_ps(in2[3], in2[3], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 e1 = _mm_shuffle_ps(in2[3], in2[3], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 e2 = _mm_shuffle_ps(in2[3], in2[3], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 e3 = _mm_shuffle_ps(in2[3], in2[3], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 m0 = _mm_mul_ps(in1[0], e0); + __m128 m1 = _mm_mul_ps(in1[1], e1); + __m128 m2 = _mm_mul_ps(in1[2], e2); + __m128 m3 = _mm_mul_ps(in1[3], e3); + + __m128 a0 = _mm_add_ps(m0, m1); + __m128 a1 = _mm_add_ps(m2, m3); + __m128 a2 = _mm_add_ps(a0, a1); + + out[3] = a2; + } + } + + GLM_FUNC_QUALIFIER void sse_transpose_ps(__m128 const in[4], __m128 out[4]) + { + __m128 tmp0 = _mm_shuffle_ps(in[0], in[1], 0x44); + __m128 tmp2 = _mm_shuffle_ps(in[0], in[1], 0xEE); + __m128 tmp1 = _mm_shuffle_ps(in[2], in[3], 0x44); + __m128 tmp3 = _mm_shuffle_ps(in[2], in[3], 0xEE); + + out[0] = _mm_shuffle_ps(tmp0, tmp1, 0x88); + out[1] = _mm_shuffle_ps(tmp0, tmp1, 0xDD); + out[2] = _mm_shuffle_ps(tmp2, tmp3, 0x88); + out[3] = _mm_shuffle_ps(tmp2, tmp3, 0xDD); + } + + GLM_FUNC_QUALIFIER __m128 sse_slow_det_ps(__m128 const in[4]) + { + __m128 Fac0; + { + // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; + // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; + // valType SubFactor06 = m[1][2] * m[3][3] - m[3][2] * m[1][3]; + // valType SubFactor13 = m[1][2] * m[2][3] - m[2][2] * m[1][3]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac0 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac1; + { + // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; + // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; + // valType SubFactor07 = m[1][1] * m[3][3] - m[3][1] * m[1][3]; + // valType SubFactor14 = m[1][1] * m[2][3] - m[2][1] * m[1][3]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac1 = _mm_sub_ps(Mul00, Mul01); + } + + + __m128 Fac2; + { + // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; + // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; + // valType SubFactor08 = m[1][1] * m[3][2] - m[3][1] * m[1][2]; + // valType SubFactor15 = m[1][1] * m[2][2] - m[2][1] * m[1][2]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac2 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac3; + { + // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; + // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; + // valType SubFactor09 = m[1][0] * m[3][3] - m[3][0] * m[1][3]; + // valType SubFactor16 = m[1][0] * m[2][3] - m[2][0] * m[1][3]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac3 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac4; + { + // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; + // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; + // valType SubFactor10 = m[1][0] * m[3][2] - m[3][0] * m[1][2]; + // valType SubFactor17 = m[1][0] * m[2][2] - m[2][0] * m[1][2]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac4 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac5; + { + // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; + // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; + // valType SubFactor12 = m[1][0] * m[3][1] - m[3][0] * m[1][1]; + // valType SubFactor18 = m[1][0] * m[2][1] - m[2][0] * m[1][1]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac5 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 SignA = _mm_set_ps( 1.0f,-1.0f, 1.0f,-1.0f); + __m128 SignB = _mm_set_ps(-1.0f, 1.0f,-1.0f, 1.0f); + + // m[1][0] + // m[0][0] + // m[0][0] + // m[0][0] + __m128 Temp0 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Vec0 = _mm_shuffle_ps(Temp0, Temp0, _MM_SHUFFLE(2, 2, 2, 0)); + + // m[1][1] + // m[0][1] + // m[0][1] + // m[0][1] + __m128 Temp1 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Vec1 = _mm_shuffle_ps(Temp1, Temp1, _MM_SHUFFLE(2, 2, 2, 0)); + + // m[1][2] + // m[0][2] + // m[0][2] + // m[0][2] + __m128 Temp2 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Vec2 = _mm_shuffle_ps(Temp2, Temp2, _MM_SHUFFLE(2, 2, 2, 0)); + + // m[1][3] + // m[0][3] + // m[0][3] + // m[0][3] + __m128 Temp3 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Vec3 = _mm_shuffle_ps(Temp3, Temp3, _MM_SHUFFLE(2, 2, 2, 0)); + + // col0 + // + (Vec1[0] * Fac0[0] - Vec2[0] * Fac1[0] + Vec3[0] * Fac2[0]), + // - (Vec1[1] * Fac0[1] - Vec2[1] * Fac1[1] + Vec3[1] * Fac2[1]), + // + (Vec1[2] * Fac0[2] - Vec2[2] * Fac1[2] + Vec3[2] * Fac2[2]), + // - (Vec1[3] * Fac0[3] - Vec2[3] * Fac1[3] + Vec3[3] * Fac2[3]), + __m128 Mul00 = _mm_mul_ps(Vec1, Fac0); + __m128 Mul01 = _mm_mul_ps(Vec2, Fac1); + __m128 Mul02 = _mm_mul_ps(Vec3, Fac2); + __m128 Sub00 = _mm_sub_ps(Mul00, Mul01); + __m128 Add00 = _mm_add_ps(Sub00, Mul02); + __m128 Inv0 = _mm_mul_ps(SignB, Add00); + + // col1 + // - (Vec0[0] * Fac0[0] - Vec2[0] * Fac3[0] + Vec3[0] * Fac4[0]), + // + (Vec0[0] * Fac0[1] - Vec2[1] * Fac3[1] + Vec3[1] * Fac4[1]), + // - (Vec0[0] * Fac0[2] - Vec2[2] * Fac3[2] + Vec3[2] * Fac4[2]), + // + (Vec0[0] * Fac0[3] - Vec2[3] * Fac3[3] + Vec3[3] * Fac4[3]), + __m128 Mul03 = _mm_mul_ps(Vec0, Fac0); + __m128 Mul04 = _mm_mul_ps(Vec2, Fac3); + __m128 Mul05 = _mm_mul_ps(Vec3, Fac4); + __m128 Sub01 = _mm_sub_ps(Mul03, Mul04); + __m128 Add01 = _mm_add_ps(Sub01, Mul05); + __m128 Inv1 = _mm_mul_ps(SignA, Add01); + + // col2 + // + (Vec0[0] * Fac1[0] - Vec1[0] * Fac3[0] + Vec3[0] * Fac5[0]), + // - (Vec0[0] * Fac1[1] - Vec1[1] * Fac3[1] + Vec3[1] * Fac5[1]), + // + (Vec0[0] * Fac1[2] - Vec1[2] * Fac3[2] + Vec3[2] * Fac5[2]), + // - (Vec0[0] * Fac1[3] - Vec1[3] * Fac3[3] + Vec3[3] * Fac5[3]), + __m128 Mul06 = _mm_mul_ps(Vec0, Fac1); + __m128 Mul07 = _mm_mul_ps(Vec1, Fac3); + __m128 Mul08 = _mm_mul_ps(Vec3, Fac5); + __m128 Sub02 = _mm_sub_ps(Mul06, Mul07); + __m128 Add02 = _mm_add_ps(Sub02, Mul08); + __m128 Inv2 = _mm_mul_ps(SignB, Add02); + + // col3 + // - (Vec1[0] * Fac2[0] - Vec1[0] * Fac4[0] + Vec2[0] * Fac5[0]), + // + (Vec1[0] * Fac2[1] - Vec1[1] * Fac4[1] + Vec2[1] * Fac5[1]), + // - (Vec1[0] * Fac2[2] - Vec1[2] * Fac4[2] + Vec2[2] * Fac5[2]), + // + (Vec1[0] * Fac2[3] - Vec1[3] * Fac4[3] + Vec2[3] * Fac5[3])); + __m128 Mul09 = _mm_mul_ps(Vec0, Fac2); + __m128 Mul10 = _mm_mul_ps(Vec1, Fac4); + __m128 Mul11 = _mm_mul_ps(Vec2, Fac5); + __m128 Sub03 = _mm_sub_ps(Mul09, Mul10); + __m128 Add03 = _mm_add_ps(Sub03, Mul11); + __m128 Inv3 = _mm_mul_ps(SignA, Add03); + + __m128 Row0 = _mm_shuffle_ps(Inv0, Inv1, _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Row1 = _mm_shuffle_ps(Inv2, Inv3, _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Row2 = _mm_shuffle_ps(Row0, Row1, _MM_SHUFFLE(2, 0, 2, 0)); + + // valType Determinant = m[0][0] * Inverse[0][0] + // + m[0][1] * Inverse[1][0] + // + m[0][2] * Inverse[2][0] + // + m[0][3] * Inverse[3][0]; + __m128 Det0 = sse_dot_ps(in[0], Row2); + return Det0; + } + + GLM_FUNC_QUALIFIER __m128 sse_detd_ps + ( + __m128 const m[4] + ) + { + // _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128( + + //T SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; + //T SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; + //T SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; + //T SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; + //T SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; + //T SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; + + // First 2 columns + __m128 Swp2A = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[2]), _MM_SHUFFLE(0, 1, 1, 2))); + __m128 Swp3A = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[3]), _MM_SHUFFLE(3, 2, 3, 3))); + __m128 MulA = _mm_mul_ps(Swp2A, Swp3A); + + // Second 2 columns + __m128 Swp2B = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[2]), _MM_SHUFFLE(3, 2, 3, 3))); + __m128 Swp3B = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[3]), _MM_SHUFFLE(0, 1, 1, 2))); + __m128 MulB = _mm_mul_ps(Swp2B, Swp3B); + + // Columns subtraction + __m128 SubE = _mm_sub_ps(MulA, MulB); + + // Last 2 rows + __m128 Swp2C = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[2]), _MM_SHUFFLE(0, 0, 1, 2))); + __m128 Swp3C = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[3]), _MM_SHUFFLE(1, 2, 0, 0))); + __m128 MulC = _mm_mul_ps(Swp2C, Swp3C); + __m128 SubF = _mm_sub_ps(_mm_movehl_ps(MulC, MulC), MulC); + + //tvec4 DetCof( + // + (m[1][1] * SubFactor00 - m[1][2] * SubFactor01 + m[1][3] * SubFactor02), + // - (m[1][0] * SubFactor00 - m[1][2] * SubFactor03 + m[1][3] * SubFactor04), + // + (m[1][0] * SubFactor01 - m[1][1] * SubFactor03 + m[1][3] * SubFactor05), + // - (m[1][0] * SubFactor02 - m[1][1] * SubFactor04 + m[1][2] * SubFactor05)); + + __m128 SubFacA = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(SubE), _MM_SHUFFLE(2, 1, 0, 0))); + __m128 SwpFacA = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[1]), _MM_SHUFFLE(0, 0, 0, 1))); + __m128 MulFacA = _mm_mul_ps(SwpFacA, SubFacA); + + __m128 SubTmpB = _mm_shuffle_ps(SubE, SubF, _MM_SHUFFLE(0, 0, 3, 1)); + __m128 SubFacB = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(SubTmpB), _MM_SHUFFLE(3, 1, 1, 0)));//SubF[0], SubE[3], SubE[3], SubE[1]; + __m128 SwpFacB = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[1]), _MM_SHUFFLE(1, 1, 2, 2))); + __m128 MulFacB = _mm_mul_ps(SwpFacB, SubFacB); + + __m128 SubRes = _mm_sub_ps(MulFacA, MulFacB); + + __m128 SubTmpC = _mm_shuffle_ps(SubE, SubF, _MM_SHUFFLE(1, 0, 2, 2)); + __m128 SubFacC = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(SubTmpC), _MM_SHUFFLE(3, 3, 2, 0))); + __m128 SwpFacC = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[1]), _MM_SHUFFLE(2, 3, 3, 3))); + __m128 MulFacC = _mm_mul_ps(SwpFacC, SubFacC); + + __m128 AddRes = _mm_add_ps(SubRes, MulFacC); + __m128 DetCof = _mm_mul_ps(AddRes, _mm_setr_ps( 1.0f,-1.0f, 1.0f,-1.0f)); + + //return m[0][0] * DetCof[0] + // + m[0][1] * DetCof[1] + // + m[0][2] * DetCof[2] + // + m[0][3] * DetCof[3]; + + return sse_dot_ps(m[0], DetCof); + } + + GLM_FUNC_QUALIFIER __m128 sse_det_ps + ( + __m128 const m[4] + ) + { + // _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(add) + + //T SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; + //T SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; + //T SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; + //T SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; + //T SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; + //T SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; + + // First 2 columns + __m128 Swp2A = _mm_shuffle_ps(m[2], m[2], _MM_SHUFFLE(0, 1, 1, 2)); + __m128 Swp3A = _mm_shuffle_ps(m[3], m[3], _MM_SHUFFLE(3, 2, 3, 3)); + __m128 MulA = _mm_mul_ps(Swp2A, Swp3A); + + // Second 2 columns + __m128 Swp2B = _mm_shuffle_ps(m[2], m[2], _MM_SHUFFLE(3, 2, 3, 3)); + __m128 Swp3B = _mm_shuffle_ps(m[3], m[3], _MM_SHUFFLE(0, 1, 1, 2)); + __m128 MulB = _mm_mul_ps(Swp2B, Swp3B); + + // Columns subtraction + __m128 SubE = _mm_sub_ps(MulA, MulB); + + // Last 2 rows + __m128 Swp2C = _mm_shuffle_ps(m[2], m[2], _MM_SHUFFLE(0, 0, 1, 2)); + __m128 Swp3C = _mm_shuffle_ps(m[3], m[3], _MM_SHUFFLE(1, 2, 0, 0)); + __m128 MulC = _mm_mul_ps(Swp2C, Swp3C); + __m128 SubF = _mm_sub_ps(_mm_movehl_ps(MulC, MulC), MulC); + + //tvec4 DetCof( + // + (m[1][1] * SubFactor00 - m[1][2] * SubFactor01 + m[1][3] * SubFactor02), + // - (m[1][0] * SubFactor00 - m[1][2] * SubFactor03 + m[1][3] * SubFactor04), + // + (m[1][0] * SubFactor01 - m[1][1] * SubFactor03 + m[1][3] * SubFactor05), + // - (m[1][0] * SubFactor02 - m[1][1] * SubFactor04 + m[1][2] * SubFactor05)); + + __m128 SubFacA = _mm_shuffle_ps(SubE, SubE, _MM_SHUFFLE(2, 1, 0, 0)); + __m128 SwpFacA = _mm_shuffle_ps(m[1], m[1], _MM_SHUFFLE(0, 0, 0, 1)); + __m128 MulFacA = _mm_mul_ps(SwpFacA, SubFacA); + + __m128 SubTmpB = _mm_shuffle_ps(SubE, SubF, _MM_SHUFFLE(0, 0, 3, 1)); + __m128 SubFacB = _mm_shuffle_ps(SubTmpB, SubTmpB, _MM_SHUFFLE(3, 1, 1, 0));//SubF[0], SubE[3], SubE[3], SubE[1]; + __m128 SwpFacB = _mm_shuffle_ps(m[1], m[1], _MM_SHUFFLE(1, 1, 2, 2)); + __m128 MulFacB = _mm_mul_ps(SwpFacB, SubFacB); + + __m128 SubRes = _mm_sub_ps(MulFacA, MulFacB); + + __m128 SubTmpC = _mm_shuffle_ps(SubE, SubF, _MM_SHUFFLE(1, 0, 2, 2)); + __m128 SubFacC = _mm_shuffle_ps(SubTmpC, SubTmpC, _MM_SHUFFLE(3, 3, 2, 0)); + __m128 SwpFacC = _mm_shuffle_ps(m[1], m[1], _MM_SHUFFLE(2, 3, 3, 3)); + __m128 MulFacC = _mm_mul_ps(SwpFacC, SubFacC); + + __m128 AddRes = _mm_add_ps(SubRes, MulFacC); + __m128 DetCof = _mm_mul_ps(AddRes, _mm_setr_ps( 1.0f,-1.0f, 1.0f,-1.0f)); + + //return m[0][0] * DetCof[0] + // + m[0][1] * DetCof[1] + // + m[0][2] * DetCof[2] + // + m[0][3] * DetCof[3]; + + return sse_dot_ps(m[0], DetCof); + } + + GLM_FUNC_QUALIFIER void sse_inverse_ps(__m128 const in[4], __m128 out[4]) + { + __m128 Fac0; + { + // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; + // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; + // valType SubFactor06 = m[1][2] * m[3][3] - m[3][2] * m[1][3]; + // valType SubFactor13 = m[1][2] * m[2][3] - m[2][2] * m[1][3]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac0 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac1; + { + // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; + // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; + // valType SubFactor07 = m[1][1] * m[3][3] - m[3][1] * m[1][3]; + // valType SubFactor14 = m[1][1] * m[2][3] - m[2][1] * m[1][3]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac1 = _mm_sub_ps(Mul00, Mul01); + } + + + __m128 Fac2; + { + // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; + // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; + // valType SubFactor08 = m[1][1] * m[3][2] - m[3][1] * m[1][2]; + // valType SubFactor15 = m[1][1] * m[2][2] - m[2][1] * m[1][2]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac2 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac3; + { + // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; + // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; + // valType SubFactor09 = m[1][0] * m[3][3] - m[3][0] * m[1][3]; + // valType SubFactor16 = m[1][0] * m[2][3] - m[2][0] * m[1][3]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac3 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac4; + { + // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; + // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; + // valType SubFactor10 = m[1][0] * m[3][2] - m[3][0] * m[1][2]; + // valType SubFactor17 = m[1][0] * m[2][2] - m[2][0] * m[1][2]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac4 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac5; + { + // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; + // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; + // valType SubFactor12 = m[1][0] * m[3][1] - m[3][0] * m[1][1]; + // valType SubFactor18 = m[1][0] * m[2][1] - m[2][0] * m[1][1]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac5 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 SignA = _mm_set_ps( 1.0f,-1.0f, 1.0f,-1.0f); + __m128 SignB = _mm_set_ps(-1.0f, 1.0f,-1.0f, 1.0f); + + // m[1][0] + // m[0][0] + // m[0][0] + // m[0][0] + __m128 Temp0 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Vec0 = _mm_shuffle_ps(Temp0, Temp0, _MM_SHUFFLE(2, 2, 2, 0)); + + // m[1][1] + // m[0][1] + // m[0][1] + // m[0][1] + __m128 Temp1 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Vec1 = _mm_shuffle_ps(Temp1, Temp1, _MM_SHUFFLE(2, 2, 2, 0)); + + // m[1][2] + // m[0][2] + // m[0][2] + // m[0][2] + __m128 Temp2 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Vec2 = _mm_shuffle_ps(Temp2, Temp2, _MM_SHUFFLE(2, 2, 2, 0)); + + // m[1][3] + // m[0][3] + // m[0][3] + // m[0][3] + __m128 Temp3 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Vec3 = _mm_shuffle_ps(Temp3, Temp3, _MM_SHUFFLE(2, 2, 2, 0)); + + // col0 + // + (Vec1[0] * Fac0[0] - Vec2[0] * Fac1[0] + Vec3[0] * Fac2[0]), + // - (Vec1[1] * Fac0[1] - Vec2[1] * Fac1[1] + Vec3[1] * Fac2[1]), + // + (Vec1[2] * Fac0[2] - Vec2[2] * Fac1[2] + Vec3[2] * Fac2[2]), + // - (Vec1[3] * Fac0[3] - Vec2[3] * Fac1[3] + Vec3[3] * Fac2[3]), + __m128 Mul00 = _mm_mul_ps(Vec1, Fac0); + __m128 Mul01 = _mm_mul_ps(Vec2, Fac1); + __m128 Mul02 = _mm_mul_ps(Vec3, Fac2); + __m128 Sub00 = _mm_sub_ps(Mul00, Mul01); + __m128 Add00 = _mm_add_ps(Sub00, Mul02); + __m128 Inv0 = _mm_mul_ps(SignB, Add00); + + // col1 + // - (Vec0[0] * Fac0[0] - Vec2[0] * Fac3[0] + Vec3[0] * Fac4[0]), + // + (Vec0[0] * Fac0[1] - Vec2[1] * Fac3[1] + Vec3[1] * Fac4[1]), + // - (Vec0[0] * Fac0[2] - Vec2[2] * Fac3[2] + Vec3[2] * Fac4[2]), + // + (Vec0[0] * Fac0[3] - Vec2[3] * Fac3[3] + Vec3[3] * Fac4[3]), + __m128 Mul03 = _mm_mul_ps(Vec0, Fac0); + __m128 Mul04 = _mm_mul_ps(Vec2, Fac3); + __m128 Mul05 = _mm_mul_ps(Vec3, Fac4); + __m128 Sub01 = _mm_sub_ps(Mul03, Mul04); + __m128 Add01 = _mm_add_ps(Sub01, Mul05); + __m128 Inv1 = _mm_mul_ps(SignA, Add01); + + // col2 + // + (Vec0[0] * Fac1[0] - Vec1[0] * Fac3[0] + Vec3[0] * Fac5[0]), + // - (Vec0[0] * Fac1[1] - Vec1[1] * Fac3[1] + Vec3[1] * Fac5[1]), + // + (Vec0[0] * Fac1[2] - Vec1[2] * Fac3[2] + Vec3[2] * Fac5[2]), + // - (Vec0[0] * Fac1[3] - Vec1[3] * Fac3[3] + Vec3[3] * Fac5[3]), + __m128 Mul06 = _mm_mul_ps(Vec0, Fac1); + __m128 Mul07 = _mm_mul_ps(Vec1, Fac3); + __m128 Mul08 = _mm_mul_ps(Vec3, Fac5); + __m128 Sub02 = _mm_sub_ps(Mul06, Mul07); + __m128 Add02 = _mm_add_ps(Sub02, Mul08); + __m128 Inv2 = _mm_mul_ps(SignB, Add02); + + // col3 + // - (Vec1[0] * Fac2[0] - Vec1[0] * Fac4[0] + Vec2[0] * Fac5[0]), + // + (Vec1[0] * Fac2[1] - Vec1[1] * Fac4[1] + Vec2[1] * Fac5[1]), + // - (Vec1[0] * Fac2[2] - Vec1[2] * Fac4[2] + Vec2[2] * Fac5[2]), + // + (Vec1[0] * Fac2[3] - Vec1[3] * Fac4[3] + Vec2[3] * Fac5[3])); + __m128 Mul09 = _mm_mul_ps(Vec0, Fac2); + __m128 Mul10 = _mm_mul_ps(Vec1, Fac4); + __m128 Mul11 = _mm_mul_ps(Vec2, Fac5); + __m128 Sub03 = _mm_sub_ps(Mul09, Mul10); + __m128 Add03 = _mm_add_ps(Sub03, Mul11); + __m128 Inv3 = _mm_mul_ps(SignA, Add03); + + __m128 Row0 = _mm_shuffle_ps(Inv0, Inv1, _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Row1 = _mm_shuffle_ps(Inv2, Inv3, _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Row2 = _mm_shuffle_ps(Row0, Row1, _MM_SHUFFLE(2, 0, 2, 0)); + + // valType Determinant = m[0][0] * Inverse[0][0] + // + m[0][1] * Inverse[1][0] + // + m[0][2] * Inverse[2][0] + // + m[0][3] * Inverse[3][0]; + __m128 Det0 = sse_dot_ps(in[0], Row2); + __m128 Rcp0 = _mm_div_ps(one, Det0); + //__m128 Rcp0 = _mm_rcp_ps(Det0); + + // Inverse /= Determinant; + out[0] = _mm_mul_ps(Inv0, Rcp0); + out[1] = _mm_mul_ps(Inv1, Rcp0); + out[2] = _mm_mul_ps(Inv2, Rcp0); + out[3] = _mm_mul_ps(Inv3, Rcp0); + } + + GLM_FUNC_QUALIFIER void sse_inverse_fast_ps(__m128 const in[4], __m128 out[4]) + { + __m128 Fac0; + { + // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; + // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; + // valType SubFactor06 = m[1][2] * m[3][3] - m[3][2] * m[1][3]; + // valType SubFactor13 = m[1][2] * m[2][3] - m[2][2] * m[1][3]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac0 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac1; + { + // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; + // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; + // valType SubFactor07 = m[1][1] * m[3][3] - m[3][1] * m[1][3]; + // valType SubFactor14 = m[1][1] * m[2][3] - m[2][1] * m[1][3]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac1 = _mm_sub_ps(Mul00, Mul01); + } + + + __m128 Fac2; + { + // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; + // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; + // valType SubFactor08 = m[1][1] * m[3][2] - m[3][1] * m[1][2]; + // valType SubFactor15 = m[1][1] * m[2][2] - m[2][1] * m[1][2]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac2 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac3; + { + // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; + // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; + // valType SubFactor09 = m[1][0] * m[3][3] - m[3][0] * m[1][3]; + // valType SubFactor16 = m[1][0] * m[2][3] - m[2][0] * m[1][3]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac3 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac4; + { + // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; + // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; + // valType SubFactor10 = m[1][0] * m[3][2] - m[3][0] * m[1][2]; + // valType SubFactor17 = m[1][0] * m[2][2] - m[2][0] * m[1][2]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac4 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 Fac5; + { + // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; + // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; + // valType SubFactor12 = m[1][0] * m[3][1] - m[3][0] * m[1][1]; + // valType SubFactor18 = m[1][0] * m[2][1] - m[2][0] * m[1][1]; + + __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); + + __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); + __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); + + __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); + __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); + Fac5 = _mm_sub_ps(Mul00, Mul01); + } + + __m128 SignA = _mm_set_ps( 1.0f,-1.0f, 1.0f,-1.0f); + __m128 SignB = _mm_set_ps(-1.0f, 1.0f,-1.0f, 1.0f); + + // m[1][0] + // m[0][0] + // m[0][0] + // m[0][0] + __m128 Temp0 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Vec0 = _mm_shuffle_ps(Temp0, Temp0, _MM_SHUFFLE(2, 2, 2, 0)); + + // m[1][1] + // m[0][1] + // m[0][1] + // m[0][1] + __m128 Temp1 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(1, 1, 1, 1)); + __m128 Vec1 = _mm_shuffle_ps(Temp1, Temp1, _MM_SHUFFLE(2, 2, 2, 0)); + + // m[1][2] + // m[0][2] + // m[0][2] + // m[0][2] + __m128 Temp2 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(2, 2, 2, 2)); + __m128 Vec2 = _mm_shuffle_ps(Temp2, Temp2, _MM_SHUFFLE(2, 2, 2, 0)); + + // m[1][3] + // m[0][3] + // m[0][3] + // m[0][3] + __m128 Temp3 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(3, 3, 3, 3)); + __m128 Vec3 = _mm_shuffle_ps(Temp3, Temp3, _MM_SHUFFLE(2, 2, 2, 0)); + + // col0 + // + (Vec1[0] * Fac0[0] - Vec2[0] * Fac1[0] + Vec3[0] * Fac2[0]), + // - (Vec1[1] * Fac0[1] - Vec2[1] * Fac1[1] + Vec3[1] * Fac2[1]), + // + (Vec1[2] * Fac0[2] - Vec2[2] * Fac1[2] + Vec3[2] * Fac2[2]), + // - (Vec1[3] * Fac0[3] - Vec2[3] * Fac1[3] + Vec3[3] * Fac2[3]), + __m128 Mul00 = _mm_mul_ps(Vec1, Fac0); + __m128 Mul01 = _mm_mul_ps(Vec2, Fac1); + __m128 Mul02 = _mm_mul_ps(Vec3, Fac2); + __m128 Sub00 = _mm_sub_ps(Mul00, Mul01); + __m128 Add00 = _mm_add_ps(Sub00, Mul02); + __m128 Inv0 = _mm_mul_ps(SignB, Add00); + + // col1 + // - (Vec0[0] * Fac0[0] - Vec2[0] * Fac3[0] + Vec3[0] * Fac4[0]), + // + (Vec0[0] * Fac0[1] - Vec2[1] * Fac3[1] + Vec3[1] * Fac4[1]), + // - (Vec0[0] * Fac0[2] - Vec2[2] * Fac3[2] + Vec3[2] * Fac4[2]), + // + (Vec0[0] * Fac0[3] - Vec2[3] * Fac3[3] + Vec3[3] * Fac4[3]), + __m128 Mul03 = _mm_mul_ps(Vec0, Fac0); + __m128 Mul04 = _mm_mul_ps(Vec2, Fac3); + __m128 Mul05 = _mm_mul_ps(Vec3, Fac4); + __m128 Sub01 = _mm_sub_ps(Mul03, Mul04); + __m128 Add01 = _mm_add_ps(Sub01, Mul05); + __m128 Inv1 = _mm_mul_ps(SignA, Add01); + + // col2 + // + (Vec0[0] * Fac1[0] - Vec1[0] * Fac3[0] + Vec3[0] * Fac5[0]), + // - (Vec0[0] * Fac1[1] - Vec1[1] * Fac3[1] + Vec3[1] * Fac5[1]), + // + (Vec0[0] * Fac1[2] - Vec1[2] * Fac3[2] + Vec3[2] * Fac5[2]), + // - (Vec0[0] * Fac1[3] - Vec1[3] * Fac3[3] + Vec3[3] * Fac5[3]), + __m128 Mul06 = _mm_mul_ps(Vec0, Fac1); + __m128 Mul07 = _mm_mul_ps(Vec1, Fac3); + __m128 Mul08 = _mm_mul_ps(Vec3, Fac5); + __m128 Sub02 = _mm_sub_ps(Mul06, Mul07); + __m128 Add02 = _mm_add_ps(Sub02, Mul08); + __m128 Inv2 = _mm_mul_ps(SignB, Add02); + + // col3 + // - (Vec1[0] * Fac2[0] - Vec1[0] * Fac4[0] + Vec2[0] * Fac5[0]), + // + (Vec1[0] * Fac2[1] - Vec1[1] * Fac4[1] + Vec2[1] * Fac5[1]), + // - (Vec1[0] * Fac2[2] - Vec1[2] * Fac4[2] + Vec2[2] * Fac5[2]), + // + (Vec1[0] * Fac2[3] - Vec1[3] * Fac4[3] + Vec2[3] * Fac5[3])); + __m128 Mul09 = _mm_mul_ps(Vec0, Fac2); + __m128 Mul10 = _mm_mul_ps(Vec1, Fac4); + __m128 Mul11 = _mm_mul_ps(Vec2, Fac5); + __m128 Sub03 = _mm_sub_ps(Mul09, Mul10); + __m128 Add03 = _mm_add_ps(Sub03, Mul11); + __m128 Inv3 = _mm_mul_ps(SignA, Add03); + + __m128 Row0 = _mm_shuffle_ps(Inv0, Inv1, _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Row1 = _mm_shuffle_ps(Inv2, Inv3, _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Row2 = _mm_shuffle_ps(Row0, Row1, _MM_SHUFFLE(2, 0, 2, 0)); + + // valType Determinant = m[0][0] * Inverse[0][0] + // + m[0][1] * Inverse[1][0] + // + m[0][2] * Inverse[2][0] + // + m[0][3] * Inverse[3][0]; + __m128 Det0 = sse_dot_ps(in[0], Row2); + __m128 Rcp0 = _mm_rcp_ps(Det0); + //__m128 Rcp0 = _mm_div_ps(one, Det0); + // Inverse /= Determinant; + out[0] = _mm_mul_ps(Inv0, Rcp0); + out[1] = _mm_mul_ps(Inv1, Rcp0); + out[2] = _mm_mul_ps(Inv2, Rcp0); + out[3] = _mm_mul_ps(Inv3, Rcp0); + } + /* + GLM_FUNC_QUALIFIER void sse_rotate_ps(__m128 const in[4], float Angle, float const v[3], __m128 out[4]) + { + float a = glm::radians(Angle); + float c = cos(a); + float s = sin(a); + + glm::vec4 AxisA(v[0], v[1], v[2], float(0)); + __m128 AxisB = _mm_set_ps(AxisA.w, AxisA.z, AxisA.y, AxisA.x); + __m128 AxisC = detail::sse_nrm_ps(AxisB); + + __m128 Cos0 = _mm_set_ss(c); + __m128 CosA = _mm_shuffle_ps(Cos0, Cos0, _MM_SHUFFLE(0, 0, 0, 0)); + __m128 Sin0 = _mm_set_ss(s); + __m128 SinA = _mm_shuffle_ps(Sin0, Sin0, _MM_SHUFFLE(0, 0, 0, 0)); + + // tvec3 temp = (valType(1) - c) * axis; + __m128 Temp0 = _mm_sub_ps(one, CosA); + __m128 Temp1 = _mm_mul_ps(Temp0, AxisC); + + //Rotate[0][0] = c + temp[0] * axis[0]; + //Rotate[0][1] = 0 + temp[0] * axis[1] + s * axis[2]; + //Rotate[0][2] = 0 + temp[0] * axis[2] - s * axis[1]; + __m128 Axis0 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(0, 0, 0, 0)); + __m128 TmpA0 = _mm_mul_ps(Axis0, AxisC); + __m128 CosA0 = _mm_shuffle_ps(Cos0, Cos0, _MM_SHUFFLE(1, 1, 1, 0)); + __m128 TmpA1 = _mm_add_ps(CosA0, TmpA0); + __m128 SinA0 = SinA;//_mm_set_ps(0.0f, s, -s, 0.0f); + __m128 TmpA2 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(3, 1, 2, 3)); + __m128 TmpA3 = _mm_mul_ps(SinA0, TmpA2); + __m128 TmpA4 = _mm_add_ps(TmpA1, TmpA3); + + //Rotate[1][0] = 0 + temp[1] * axis[0] - s * axis[2]; + //Rotate[1][1] = c + temp[1] * axis[1]; + //Rotate[1][2] = 0 + temp[1] * axis[2] + s * axis[0]; + __m128 Axis1 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(1, 1, 1, 1)); + __m128 TmpB0 = _mm_mul_ps(Axis1, AxisC); + __m128 CosA1 = _mm_shuffle_ps(Cos0, Cos0, _MM_SHUFFLE(1, 1, 0, 1)); + __m128 TmpB1 = _mm_add_ps(CosA1, TmpB0); + __m128 SinB0 = SinA;//_mm_set_ps(-s, 0.0f, s, 0.0f); + __m128 TmpB2 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(3, 0, 3, 2)); + __m128 TmpB3 = _mm_mul_ps(SinA0, TmpB2); + __m128 TmpB4 = _mm_add_ps(TmpB1, TmpB3); + + //Rotate[2][0] = 0 + temp[2] * axis[0] + s * axis[1]; + //Rotate[2][1] = 0 + temp[2] * axis[1] - s * axis[0]; + //Rotate[2][2] = c + temp[2] * axis[2]; + __m128 Axis2 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(2, 2, 2, 2)); + __m128 TmpC0 = _mm_mul_ps(Axis2, AxisC); + __m128 CosA2 = _mm_shuffle_ps(Cos0, Cos0, _MM_SHUFFLE(1, 0, 1, 1)); + __m128 TmpC1 = _mm_add_ps(CosA2, TmpC0); + __m128 SinC0 = SinA;//_mm_set_ps(s, -s, 0.0f, 0.0f); + __m128 TmpC2 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(3, 3, 0, 1)); + __m128 TmpC3 = _mm_mul_ps(SinA0, TmpC2); + __m128 TmpC4 = _mm_add_ps(TmpC1, TmpC3); + + __m128 Result[4]; + Result[0] = TmpA4; + Result[1] = TmpB4; + Result[2] = TmpC4; + Result[3] = _mm_set_ps(1, 0, 0, 0); + + //tmat4x4 Result(uninitialize); + //Result[0] = m[0] * Rotate[0][0] + m[1] * Rotate[0][1] + m[2] * Rotate[0][2]; + //Result[1] = m[0] * Rotate[1][0] + m[1] * Rotate[1][1] + m[2] * Rotate[1][2]; + //Result[2] = m[0] * Rotate[2][0] + m[1] * Rotate[2][1] + m[2] * Rotate[2][2]; + //Result[3] = m[3]; + //return Result; + sse_mul_ps(in, Result, out); + } + */ + GLM_FUNC_QUALIFIER void sse_outer_ps(__m128 const & c, __m128 const & r, __m128 out[4]) + { + out[0] = _mm_mul_ps(c, _mm_shuffle_ps(r, r, _MM_SHUFFLE(0, 0, 0, 0))); + out[1] = _mm_mul_ps(c, _mm_shuffle_ps(r, r, _MM_SHUFFLE(1, 1, 1, 1))); + out[2] = _mm_mul_ps(c, _mm_shuffle_ps(r, r, _MM_SHUFFLE(2, 2, 2, 2))); + out[3] = _mm_mul_ps(c, _mm_shuffle_ps(r, r, _MM_SHUFFLE(3, 3, 3, 3))); + } +}//namespace detail + template <> GLM_FUNC_QUALIFIER tmat4x4 inverse(tmat4x4 const& m) { diff --git a/glm/detail/intrinsic_common.hpp b/glm/detail/intrinsic_common.hpp deleted file mode 100644 index 8cb1a1c6..00000000 --- a/glm/detail/intrinsic_common.hpp +++ /dev/null @@ -1,87 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_common.hpp -/// @date 2009-05-11 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -#pragma once - -#include "setup.hpp" - -#if(!(GLM_ARCH & GLM_ARCH_SSE2)) -# error "SSE2 instructions not supported or enabled" -#else - -namespace glm{ -namespace detail -{ - __m128 sse_abs_ps(__m128 x); - - __m128 sse_sgn_ps(__m128 x); - - //floor - __m128 sse_flr_ps(__m128 v); - - //trunc - __m128 sse_trc_ps(__m128 v); - - //round - __m128 sse_nd_ps(__m128 v); - - //roundEven - __m128 sse_rde_ps(__m128 v); - - __m128 sse_rnd_ps(__m128 x); - - __m128 sse_ceil_ps(__m128 v); - - __m128 sse_frc_ps(__m128 x); - - __m128 sse_mod_ps(__m128 x, __m128 y); - - __m128 sse_modf_ps(__m128 x, __m128i & i); - - //GLM_FUNC_QUALIFIER __m128 sse_min_ps(__m128 x, __m128 y) - - //GLM_FUNC_QUALIFIER __m128 sse_max_ps(__m128 x, __m128 y) - - __m128 sse_clp_ps(__m128 v, __m128 minVal, __m128 maxVal); - - __m128 sse_mix_ps(__m128 v1, __m128 v2, __m128 a); - - __m128 sse_stp_ps(__m128 edge, __m128 x); - - __m128 sse_ssp_ps(__m128 edge0, __m128 edge1, __m128 x); - - __m128 sse_nan_ps(__m128 x); - - __m128 sse_inf_ps(__m128 x); - -}//namespace detail -}//namespace glm - -#include "intrinsic_common.inl" - -#endif//GLM_ARCH diff --git a/glm/detail/intrinsic_common.inl b/glm/detail/intrinsic_common.inl deleted file mode 100644 index f34bf939..00000000 --- a/glm/detail/intrinsic_common.inl +++ /dev/null @@ -1,137 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_common.inl -/// @date 2009-05-08 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -namespace glm{ -namespace detail -{ -/* - static const __m128 GLM_VAR_USED zero = _mm_setzero_ps(); - static const __m128 GLM_VAR_USED one = _mm_set_ps1(1.0f); - static const __m128 GLM_VAR_USED minus_one = _mm_set_ps1(-1.0f); - static const __m128 GLM_VAR_USED two = _mm_set_ps1(2.0f); - static const __m128 GLM_VAR_USED three = _mm_set_ps1(3.0f); - static const __m128 GLM_VAR_USED pi = _mm_set_ps1(3.1415926535897932384626433832795f); - static const __m128 GLM_VAR_USED hundred_eighty = _mm_set_ps1(180.f); - static const __m128 GLM_VAR_USED pi_over_hundred_eighty = _mm_set_ps1(0.017453292519943295769236907684886f); - static const __m128 GLM_VAR_USED hundred_eighty_over_pi = _mm_set_ps1(57.295779513082320876798154814105f); - - static const ieee754_QNAN absMask; - static const __m128 GLM_VAR_USED abs4Mask = _mm_set_ps1(absMask.f); - - static const __m128 GLM_VAR_USED _epi32_sign_mask = _mm_castsi128_ps(_mm_set1_epi32(static_cast(0x80000000))); - //static const __m128 GLM_VAR_USED _epi32_inv_sign_mask = _mm_castsi128_ps(_mm_set1_epi32(0x7FFFFFFF)); - //static const __m128 GLM_VAR_USED _epi32_mant_mask = _mm_castsi128_ps(_mm_set1_epi32(0x7F800000)); - //static const __m128 GLM_VAR_USED _epi32_inv_mant_mask = _mm_castsi128_ps(_mm_set1_epi32(0x807FFFFF)); - //static const __m128 GLM_VAR_USED _epi32_min_norm_pos = _mm_castsi128_ps(_mm_set1_epi32(0x00800000)); - static const __m128 GLM_VAR_USED _epi32_0 = _mm_set_ps1(0); - static const __m128 GLM_VAR_USED _epi32_1 = _mm_set_ps1(1); - static const __m128 GLM_VAR_USED _epi32_2 = _mm_set_ps1(2); - static const __m128 GLM_VAR_USED _epi32_3 = _mm_set_ps1(3); - static const __m128 GLM_VAR_USED _epi32_4 = _mm_set_ps1(4); - static const __m128 GLM_VAR_USED _epi32_5 = _mm_set_ps1(5); - static const __m128 GLM_VAR_USED _epi32_6 = _mm_set_ps1(6); - static const __m128 GLM_VAR_USED _epi32_7 = _mm_set_ps1(7); - static const __m128 GLM_VAR_USED _epi32_8 = _mm_set_ps1(8); - static const __m128 GLM_VAR_USED _epi32_9 = _mm_set_ps1(9); - static const __m128 GLM_VAR_USED _epi32_127 = _mm_set_ps1(127); - //static const __m128 GLM_VAR_USED _epi32_ninf = _mm_castsi128_ps(_mm_set1_epi32(0xFF800000)); - //static const __m128 GLM_VAR_USED _epi32_pinf = _mm_castsi128_ps(_mm_set1_epi32(0x7F800000)); - - static const __m128 GLM_VAR_USED _ps_1_3 = _mm_set_ps1(0.33333333333333333333333333333333f); - static const __m128 GLM_VAR_USED _ps_0p5 = _mm_set_ps1(0.5f); - static const __m128 GLM_VAR_USED _ps_1 = _mm_set_ps1(1.0f); - static const __m128 GLM_VAR_USED _ps_m1 = _mm_set_ps1(-1.0f); - static const __m128 GLM_VAR_USED _ps_2 = _mm_set_ps1(2.0f); - static const __m128 GLM_VAR_USED _ps_3 = _mm_set_ps1(3.0f); - static const __m128 GLM_VAR_USED _ps_127 = _mm_set_ps1(127.0f); - static const __m128 GLM_VAR_USED _ps_255 = _mm_set_ps1(255.0f); - static const __m128 GLM_VAR_USED _ps_2pow23 = _mm_set_ps1(8388608.0f); - - static const __m128 GLM_VAR_USED _ps_1_0_0_0 = _mm_set_ps(1.0f, 0.0f, 0.0f, 0.0f); - static const __m128 GLM_VAR_USED _ps_0_1_0_0 = _mm_set_ps(0.0f, 1.0f, 0.0f, 0.0f); - static const __m128 GLM_VAR_USED _ps_0_0_1_0 = _mm_set_ps(0.0f, 0.0f, 1.0f, 0.0f); - static const __m128 GLM_VAR_USED _ps_0_0_0_1 = _mm_set_ps(0.0f, 0.0f, 0.0f, 1.0f); - - static const __m128 GLM_VAR_USED _ps_pi = _mm_set_ps1(3.1415926535897932384626433832795f); - static const __m128 GLM_VAR_USED _ps_pi2 = _mm_set_ps1(6.283185307179586476925286766560f); - static const __m128 GLM_VAR_USED _ps_2_pi = _mm_set_ps1(0.63661977236758134307553505349006f); - static const __m128 GLM_VAR_USED _ps_pi_2 = _mm_set_ps1(1.5707963267948966192313216916398f); - static const __m128 GLM_VAR_USED _ps_4_pi = _mm_set_ps1(1.2732395447351626861510701069801f); - static const __m128 GLM_VAR_USED _ps_pi_4 = _mm_set_ps1(0.78539816339744830961566084581988f); - - static const __m128 GLM_VAR_USED _ps_sincos_p0 = _mm_set_ps1(0.15707963267948963959e1f); - static const __m128 GLM_VAR_USED _ps_sincos_p1 = _mm_set_ps1(-0.64596409750621907082e0f); - static const __m128 GLM_VAR_USED _ps_sincos_p2 = _mm_set_ps1(0.7969262624561800806e-1f); - static const __m128 GLM_VAR_USED _ps_sincos_p3 = _mm_set_ps1(-0.468175413106023168e-2f); - static const __m128 GLM_VAR_USED _ps_tan_p0 = _mm_set_ps1(-1.79565251976484877988e7f); - static const __m128 GLM_VAR_USED _ps_tan_p1 = _mm_set_ps1(1.15351664838587416140e6f); - static const __m128 GLM_VAR_USED _ps_tan_p2 = _mm_set_ps1(-1.30936939181383777646e4f); - static const __m128 GLM_VAR_USED _ps_tan_q0 = _mm_set_ps1(-5.38695755929454629881e7f); - static const __m128 GLM_VAR_USED _ps_tan_q1 = _mm_set_ps1(2.50083801823357915839e7f); - static const __m128 GLM_VAR_USED _ps_tan_q2 = _mm_set_ps1(-1.32089234440210967447e6f); - static const __m128 GLM_VAR_USED _ps_tan_q3 = _mm_set_ps1(1.36812963470692954678e4f); - static const __m128 GLM_VAR_USED _ps_tan_poleval = _mm_set_ps1(3.68935e19f); - static const __m128 GLM_VAR_USED _ps_atan_t0 = _mm_set_ps1(-0.91646118527267623468e-1f); - static const __m128 GLM_VAR_USED _ps_atan_t1 = _mm_set_ps1(-0.13956945682312098640e1f); - static const __m128 GLM_VAR_USED _ps_atan_t2 = _mm_set_ps1(-0.94393926122725531747e2f); - static const __m128 GLM_VAR_USED _ps_atan_t3 = _mm_set_ps1(0.12888383034157279340e2f); - static const __m128 GLM_VAR_USED _ps_atan_s0 = _mm_set_ps1(0.12797564625607904396e1f); - static const __m128 GLM_VAR_USED _ps_atan_s1 = _mm_set_ps1(0.21972168858277355914e1f); - static const __m128 GLM_VAR_USED _ps_atan_s2 = _mm_set_ps1(0.68193064729268275701e1f); - static const __m128 GLM_VAR_USED _ps_atan_s3 = _mm_set_ps1(0.28205206687035841409e2f); - - static const __m128 GLM_VAR_USED _ps_exp_hi = _mm_set_ps1(88.3762626647949f); - static const __m128 GLM_VAR_USED _ps_exp_lo = _mm_set_ps1(-88.3762626647949f); - static const __m128 GLM_VAR_USED _ps_exp_rln2 = _mm_set_ps1(1.4426950408889634073599f); - static const __m128 GLM_VAR_USED _ps_exp_p0 = _mm_set_ps1(1.26177193074810590878e-4f); - static const __m128 GLM_VAR_USED _ps_exp_p1 = _mm_set_ps1(3.02994407707441961300e-2f); - static const __m128 GLM_VAR_USED _ps_exp_q0 = _mm_set_ps1(3.00198505138664455042e-6f); - static const __m128 GLM_VAR_USED _ps_exp_q1 = _mm_set_ps1(2.52448340349684104192e-3f); - static const __m128 GLM_VAR_USED _ps_exp_q2 = _mm_set_ps1(2.27265548208155028766e-1f); - static const __m128 GLM_VAR_USED _ps_exp_q3 = _mm_set_ps1(2.00000000000000000009e0f); - static const __m128 GLM_VAR_USED _ps_exp_c1 = _mm_set_ps1(6.93145751953125e-1f); - static const __m128 GLM_VAR_USED _ps_exp_c2 = _mm_set_ps1(1.42860682030941723212e-6f); - static const __m128 GLM_VAR_USED _ps_exp2_hi = _mm_set_ps1(127.4999961853f); - static const __m128 GLM_VAR_USED _ps_exp2_lo = _mm_set_ps1(-127.4999961853f); - static const __m128 GLM_VAR_USED _ps_exp2_p0 = _mm_set_ps1(2.30933477057345225087e-2f); - static const __m128 GLM_VAR_USED _ps_exp2_p1 = _mm_set_ps1(2.02020656693165307700e1f); - static const __m128 GLM_VAR_USED _ps_exp2_p2 = _mm_set_ps1(1.51390680115615096133e3f); - static const __m128 GLM_VAR_USED _ps_exp2_q0 = _mm_set_ps1(2.33184211722314911771e2f); - static const __m128 GLM_VAR_USED _ps_exp2_q1 = _mm_set_ps1(4.36821166879210612817e3f); - static const __m128 GLM_VAR_USED _ps_log_p0 = _mm_set_ps1(-7.89580278884799154124e-1f); - static const __m128 GLM_VAR_USED _ps_log_p1 = _mm_set_ps1(1.63866645699558079767e1f); - static const __m128 GLM_VAR_USED _ps_log_p2 = _mm_set_ps1(-6.41409952958715622951e1f); - static const __m128 GLM_VAR_USED _ps_log_q0 = _mm_set_ps1(-3.56722798256324312549e1f); - static const __m128 GLM_VAR_USED _ps_log_q1 = _mm_set_ps1(3.12093766372244180303e2f); - static const __m128 GLM_VAR_USED _ps_log_q2 = _mm_set_ps1(-7.69691943550460008604e2f); - static const __m128 GLM_VAR_USED _ps_log_c0 = _mm_set_ps1(0.693147180559945f); - static const __m128 GLM_VAR_USED _ps_log2_c0 = _mm_set_ps1(1.44269504088896340735992f); -*/ -}//namespace detail -}//namespace glms - diff --git a/glm/detail/intrinsic_exponential.hpp b/glm/detail/intrinsic_exponential.hpp deleted file mode 100644 index 96d0129c..00000000 --- a/glm/detail/intrinsic_exponential.hpp +++ /dev/null @@ -1,77 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_exponential.hpp -/// @date 2009-05-11 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -#pragma once - -#include "setup.hpp" - -#if(!(GLM_ARCH & GLM_ARCH_SSE2)) -# error "SSE2 instructions not supported or enabled" -#else - -namespace glm{ -namespace detail -{ -/* -GLM_FUNC_QUALIFIER __m128 sse_rsqrt_nr_ss(__m128 const x) -{ - __m128 recip = _mm_rsqrt_ss( x ); // "estimate" opcode - const static __m128 three = { 3, 3, 3, 3 }; // aligned consts for fast load - const static __m128 half = { 0.5,0.5,0.5,0.5 }; - __m128 halfrecip = _mm_mul_ss( half, recip ); - __m128 threeminus_xrr = _mm_sub_ss( three, _mm_mul_ss( x, _mm_mul_ss ( recip, recip ) ) ); - return _mm_mul_ss( halfrecip, threeminus_xrr ); -} - -GLM_FUNC_QUALIFIER __m128 sse_normalize_fast_ps( float * RESTRICT vOut, float * RESTRICT vIn ) -{ - __m128 x = _mm_load_ss(&vIn[0]); - __m128 y = _mm_load_ss(&vIn[1]); - __m128 z = _mm_load_ss(&vIn[2]); - - const __m128 l = // compute x*x + y*y + z*z - _mm_add_ss( - _mm_add_ss( _mm_mul_ss(x,x), - _mm_mul_ss(y,y) - ), - _mm_mul_ss( z, z ) - ); - - - const __m128 rsqt = _mm_rsqrt_nr_ss( l ); - _mm_store_ss( &vOut[0] , _mm_mul_ss( rsqt, x ) ); - _mm_store_ss( &vOut[1] , _mm_mul_ss( rsqt, y ) ); - _mm_store_ss( &vOut[2] , _mm_mul_ss( rsqt, z ) ); - - return _mm_mul_ss( l , rsqt ); -} -*/ -}//namespace detail -}//namespace glm - -#endif//GLM_ARCH diff --git a/glm/detail/intrinsic_exponential.inl b/glm/detail/intrinsic_exponential.inl deleted file mode 100644 index 9dba23cd..00000000 --- a/glm/detail/intrinsic_exponential.inl +++ /dev/null @@ -1,27 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_exponential.inl -/// @date 2011-06-15 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// diff --git a/glm/detail/intrinsic_geometric.hpp b/glm/detail/intrinsic_geometric.hpp deleted file mode 100644 index ecfbb032..00000000 --- a/glm/detail/intrinsic_geometric.hpp +++ /dev/null @@ -1,74 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_geometric.hpp -/// @date 2009-05-08 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -#pragma once - -#include "setup.hpp" - -#if(!(GLM_ARCH & GLM_ARCH_SSE2)) -# error "SSE2 instructions not supported or enabled" -#else - -#include "intrinsic_common.hpp" - -namespace glm{ -namespace detail -{ - //length - __m128 sse_len_ps(__m128 x); - - //distance - __m128 sse_dst_ps(__m128 p0, __m128 p1); - - //dot - __m128 sse_dot_ps(__m128 v1, __m128 v2); - - // SSE1 - __m128 sse_dot_ss(__m128 v1, __m128 v2); - - //cross - __m128 sse_xpd_ps(__m128 v1, __m128 v2); - - //normalize - __m128 sse_nrm_ps(__m128 v); - - //faceforward - __m128 sse_ffd_ps(__m128 N, __m128 I, __m128 Nref); - - //reflect - __m128 sse_rfe_ps(__m128 I, __m128 N); - - //refract - __m128 sse_rfa_ps(__m128 I, __m128 N, __m128 eta); - -}//namespace detail -}//namespace glm - -#include "intrinsic_geometric.inl" - -#endif//GLM_ARCH diff --git a/glm/detail/intrinsic_geometric.inl b/glm/detail/intrinsic_geometric.inl deleted file mode 100644 index c6183455..00000000 --- a/glm/detail/intrinsic_geometric.inl +++ /dev/null @@ -1,149 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_geometric.inl -/// @date 2009-05-08 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -/* -namespace glm{ -namespace detail{ - -//length -GLM_FUNC_QUALIFIER __m128 sse_len_ps(__m128 x) -{ - __m128 dot0 = sse_dot_ps(x, x); - __m128 sqt0 = _mm_sqrt_ps(dot0); - return sqt0; -} - -//distance -GLM_FUNC_QUALIFIER __m128 sse_dst_ps(__m128 p0, __m128 p1) -{ - __m128 sub0 = _mm_sub_ps(p0, p1); - __m128 len0 = sse_len_ps(sub0); - return len0; -} - -//dot -GLM_FUNC_QUALIFIER __m128 sse_dot_ps(__m128 v1, __m128 v2) -{ -# if(GLM_ARCH & GLM_ARCH_AVX) - return _mm_dp_ps(v1, v2, 0xff); -# else - __m128 mul0 = _mm_mul_ps(v1, v2); - __m128 swp0 = _mm_shuffle_ps(mul0, mul0, _MM_SHUFFLE(2, 3, 0, 1)); - __m128 add0 = _mm_add_ps(mul0, swp0); - __m128 swp1 = _mm_shuffle_ps(add0, add0, _MM_SHUFFLE(0, 1, 2, 3)); - __m128 add1 = _mm_add_ps(add0, swp1); - return add1; -# endif -} - -// SSE1 -GLM_FUNC_QUALIFIER __m128 sse_dot_ss(__m128 v1, __m128 v2) -{ - __m128 mul0 = _mm_mul_ps(v1, v2); - __m128 mov0 = _mm_movehl_ps(mul0, mul0); - __m128 add0 = _mm_add_ps(mov0, mul0); - __m128 swp1 = _mm_shuffle_ps(add0, add0, 1); - __m128 add1 = _mm_add_ss(add0, swp1); - return add1; -} - -//cross -GLM_FUNC_QUALIFIER __m128 sse_xpd_ps(__m128 v1, __m128 v2) -{ - __m128 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 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 mul0 = _mm_mul_ps(swp0, swp3); - __m128 mul1 = _mm_mul_ps(swp1, swp2); - __m128 sub0 = _mm_sub_ps(mul0, mul1); - return sub0; -} - -//normalize -GLM_FUNC_QUALIFIER __m128 sse_nrm_ps(__m128 v) -{ - __m128 dot0 = sse_dot_ps(v, v); - __m128 isr0 = _mm_rsqrt_ps(dot0); - __m128 mul0 = _mm_mul_ps(v, isr0); - return mul0; -} - -//faceforward -GLM_FUNC_QUALIFIER __m128 sse_ffd_ps(__m128 N, __m128 I, __m128 Nref) -{ - //__m128 dot0 = _mm_dot_ps(v, v); - //__m128 neg0 = _mm_neg_ps(N); - //__m128 sgn0 = _mm_sgn_ps(dot0); - //__m128 mix0 = _mm_mix_ps(N, neg0, sgn0); - //return mix0; - - __m128 dot0 = sse_dot_ps(Nref, I); - __m128 sgn0 = sse_sgn_ps(dot0); - __m128 mul0 = _mm_mul_ps(sgn0, glm::detail::minus_one); - __m128 mul1 = _mm_mul_ps(N, mul0); - return mul1; -} - -//reflect -GLM_FUNC_QUALIFIER __m128 sse_rfe_ps(__m128 I, __m128 N) -{ - __m128 dot0 = sse_dot_ps(N, I); - __m128 mul0 = _mm_mul_ps(N, dot0); - __m128 mul1 = _mm_mul_ps(mul0, glm::detail::two); - __m128 sub0 = _mm_sub_ps(I, mul1); - return sub0; -} - -//refract -GLM_FUNC_QUALIFIER __m128 sse_rfa_ps(__m128 I, __m128 N, __m128 eta) -{ - __m128 dot0 = sse_dot_ps(N, I); - __m128 mul0 = _mm_mul_ps(eta, eta); - __m128 mul1 = _mm_mul_ps(dot0, dot0); - __m128 sub0 = _mm_sub_ps(glm::detail::one, mul0); - __m128 sub1 = _mm_sub_ps(glm::detail::one, mul1); - __m128 mul2 = _mm_mul_ps(sub0, sub1); - - if(_mm_movemask_ps(_mm_cmplt_ss(mul2, glm::detail::zero)) == 0) - return glm::detail::zero; - - __m128 sqt0 = _mm_sqrt_ps(mul2); - __m128 mul3 = _mm_mul_ps(eta, dot0); - __m128 add0 = _mm_add_ps(mul3, sqt0); - __m128 mul4 = _mm_mul_ps(add0, N); - __m128 mul5 = _mm_mul_ps(eta, I); - __m128 sub2 = _mm_sub_ps(mul5, mul4); - - return sub2; -} - -}//namespace detail -}//namespace glm -*/ - diff --git a/glm/detail/intrinsic_integer.hpp b/glm/detail/intrinsic_integer.hpp deleted file mode 100644 index e4f07bc5..00000000 --- a/glm/detail/intrinsic_integer.hpp +++ /dev/null @@ -1,48 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2012 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_integer.hpp -/// @date 2009-05-11 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -#pragma once - -#include "glm/glm.hpp" - -#if(!(GLM_ARCH & GLM_ARCH_SSE2)) -# error "SSE2 instructions not supported or enabled" -#else - -namespace glm{ -namespace detail -{ - __m128i _mm_bit_interleave_si128(__m128i x); - __m128i _mm_bit_interleave_si128(__m128i x, __m128i y); - -}//namespace detail -}//namespace glm - -#include "intrinsic_integer.inl" - -#endif//GLM_ARCH diff --git a/glm/detail/intrinsic_matrix.hpp b/glm/detail/intrinsic_matrix.hpp deleted file mode 100644 index e41b162c..00000000 --- a/glm/detail/intrinsic_matrix.hpp +++ /dev/null @@ -1,67 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_common.hpp -/// @date 2009-06-05 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -#pragma once - -#include "setup.hpp" - -#if(!(GLM_ARCH & GLM_ARCH_SSE2)) -# error "SSE2 instructions not supported or enabled" -#else - -#include "intrinsic_geometric.hpp" - -namespace glm{ -namespace detail -{ - void sse_add_ps(__m128 in1[4], __m128 in2[4], __m128 out[4]); - - void sse_sub_ps(__m128 in1[4], __m128 in2[4], __m128 out[4]); - - __m128 sse_mul_ps(__m128 m[4], __m128 v); - - __m128 sse_mul_ps(__m128 v, __m128 m[4]); - - void sse_mul_ps(__m128 const in1[4], __m128 const in2[4], __m128 out[4]); - - void sse_transpose_ps(__m128 const in[4], __m128 out[4]); - - void sse_inverse_ps(__m128 const in[4], __m128 out[4]); - - void sse_rotate_ps(__m128 const in[4], float Angle, float const v[3], __m128 out[4]); - - __m128 sse_det_ps(__m128 const m[4]); - - __m128 sse_slow_det_ps(__m128 const m[4]); - -}//namespace detail -}//namespace glm - -#include "intrinsic_matrix.inl" - -#endif//GLM_ARCH diff --git a/glm/detail/intrinsic_matrix.inl b/glm/detail/intrinsic_matrix.inl deleted file mode 100644 index 491b58b0..00000000 --- a/glm/detail/intrinsic_matrix.inl +++ /dev/null @@ -1,1070 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_common.inl -/// @date 2009-06-05 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -namespace glm{ -namespace detail{ - -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); - -template -GLM_FUNC_QUALIFIER matType sse_comp_mul_ps -( - __m128 const in1[4], - __m128 const in2[4], - __m128 out[4] -) -{ - out[0] = _mm_mul_ps(in1[0], in2[0]); - out[1] = _mm_mul_ps(in1[1], in2[1]); - out[2] = _mm_mul_ps(in1[2], in2[2]); - out[3] = _mm_mul_ps(in1[3], in2[3]); -} - -GLM_FUNC_QUALIFIER void sse_add_ps(__m128 const in1[4], __m128 const in2[4], __m128 out[4]) -{ - { - out[0] = _mm_add_ps(in1[0], in2[0]); - out[1] = _mm_add_ps(in1[1], in2[1]); - out[2] = _mm_add_ps(in1[2], in2[2]); - out[3] = _mm_add_ps(in1[3], in2[3]); - } -} - -GLM_FUNC_QUALIFIER void sse_sub_ps(__m128 const in1[4], __m128 const in2[4], __m128 out[4]) -{ - { - out[0] = _mm_sub_ps(in1[0], in2[0]); - out[1] = _mm_sub_ps(in1[1], in2[1]); - out[2] = _mm_sub_ps(in1[2], in2[2]); - out[3] = _mm_sub_ps(in1[3], in2[3]); - } -} - -GLM_FUNC_QUALIFIER __m128 sse_mul_ps(__m128 const m[4], __m128 v) -{ - __m128 v0 = _mm_shuffle_ps(v, v, _MM_SHUFFLE(0, 0, 0, 0)); - __m128 v1 = _mm_shuffle_ps(v, v, _MM_SHUFFLE(1, 1, 1, 1)); - __m128 v2 = _mm_shuffle_ps(v, v, _MM_SHUFFLE(2, 2, 2, 2)); - __m128 v3 = _mm_shuffle_ps(v, v, _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 m0 = _mm_mul_ps(m[0], v0); - __m128 m1 = _mm_mul_ps(m[1], v1); - __m128 m2 = _mm_mul_ps(m[2], v2); - __m128 m3 = _mm_mul_ps(m[3], v3); - - __m128 a0 = _mm_add_ps(m0, m1); - __m128 a1 = _mm_add_ps(m2, m3); - __m128 a2 = _mm_add_ps(a0, a1); - - return a2; -} - -GLM_FUNC_QUALIFIER __m128 sse_mul_ps(__m128 v, __m128 const m[4]) -{ - __m128 i0 = m[0]; - __m128 i1 = m[1]; - __m128 i2 = m[2]; - __m128 i3 = m[3]; - - __m128 m0 = _mm_mul_ps(v, i0); - __m128 m1 = _mm_mul_ps(v, i1); - __m128 m2 = _mm_mul_ps(v, i2); - __m128 m3 = _mm_mul_ps(v, i3); - - __m128 u0 = _mm_unpacklo_ps(m0, m1); - __m128 u1 = _mm_unpackhi_ps(m0, m1); - __m128 a0 = _mm_add_ps(u0, u1); - - __m128 u2 = _mm_unpacklo_ps(m2, m3); - __m128 u3 = _mm_unpackhi_ps(m2, m3); - __m128 a1 = _mm_add_ps(u2, u3); - - __m128 f0 = _mm_movelh_ps(a0, a1); - __m128 f1 = _mm_movehl_ps(a1, a0); - __m128 f2 = _mm_add_ps(f0, f1); - - return f2; -} - -GLM_FUNC_QUALIFIER void sse_mul_ps(__m128 const in1[4], __m128 const in2[4], __m128 out[4]) -{ - { - __m128 e0 = _mm_shuffle_ps(in2[0], in2[0], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 e1 = _mm_shuffle_ps(in2[0], in2[0], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 e2 = _mm_shuffle_ps(in2[0], in2[0], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 e3 = _mm_shuffle_ps(in2[0], in2[0], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 m0 = _mm_mul_ps(in1[0], e0); - __m128 m1 = _mm_mul_ps(in1[1], e1); - __m128 m2 = _mm_mul_ps(in1[2], e2); - __m128 m3 = _mm_mul_ps(in1[3], e3); - - __m128 a0 = _mm_add_ps(m0, m1); - __m128 a1 = _mm_add_ps(m2, m3); - __m128 a2 = _mm_add_ps(a0, a1); - - out[0] = a2; - } - - { - __m128 e0 = _mm_shuffle_ps(in2[1], in2[1], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 e1 = _mm_shuffle_ps(in2[1], in2[1], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 e2 = _mm_shuffle_ps(in2[1], in2[1], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 e3 = _mm_shuffle_ps(in2[1], in2[1], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 m0 = _mm_mul_ps(in1[0], e0); - __m128 m1 = _mm_mul_ps(in1[1], e1); - __m128 m2 = _mm_mul_ps(in1[2], e2); - __m128 m3 = _mm_mul_ps(in1[3], e3); - - __m128 a0 = _mm_add_ps(m0, m1); - __m128 a1 = _mm_add_ps(m2, m3); - __m128 a2 = _mm_add_ps(a0, a1); - - out[1] = a2; - } - - { - __m128 e0 = _mm_shuffle_ps(in2[2], in2[2], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 e1 = _mm_shuffle_ps(in2[2], in2[2], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 e2 = _mm_shuffle_ps(in2[2], in2[2], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 e3 = _mm_shuffle_ps(in2[2], in2[2], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 m0 = _mm_mul_ps(in1[0], e0); - __m128 m1 = _mm_mul_ps(in1[1], e1); - __m128 m2 = _mm_mul_ps(in1[2], e2); - __m128 m3 = _mm_mul_ps(in1[3], e3); - - __m128 a0 = _mm_add_ps(m0, m1); - __m128 a1 = _mm_add_ps(m2, m3); - __m128 a2 = _mm_add_ps(a0, a1); - - out[2] = a2; - } - - { - //(__m128&)_mm_shuffle_epi32(__m128i&)in2[0], _MM_SHUFFLE(3, 3, 3, 3)) - __m128 e0 = _mm_shuffle_ps(in2[3], in2[3], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 e1 = _mm_shuffle_ps(in2[3], in2[3], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 e2 = _mm_shuffle_ps(in2[3], in2[3], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 e3 = _mm_shuffle_ps(in2[3], in2[3], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 m0 = _mm_mul_ps(in1[0], e0); - __m128 m1 = _mm_mul_ps(in1[1], e1); - __m128 m2 = _mm_mul_ps(in1[2], e2); - __m128 m3 = _mm_mul_ps(in1[3], e3); - - __m128 a0 = _mm_add_ps(m0, m1); - __m128 a1 = _mm_add_ps(m2, m3); - __m128 a2 = _mm_add_ps(a0, a1); - - out[3] = a2; - } -} - -GLM_FUNC_QUALIFIER void sse_transpose_ps(__m128 const in[4], __m128 out[4]) -{ - __m128 tmp0 = _mm_shuffle_ps(in[0], in[1], 0x44); - __m128 tmp2 = _mm_shuffle_ps(in[0], in[1], 0xEE); - __m128 tmp1 = _mm_shuffle_ps(in[2], in[3], 0x44); - __m128 tmp3 = _mm_shuffle_ps(in[2], in[3], 0xEE); - - out[0] = _mm_shuffle_ps(tmp0, tmp1, 0x88); - out[1] = _mm_shuffle_ps(tmp0, tmp1, 0xDD); - out[2] = _mm_shuffle_ps(tmp2, tmp3, 0x88); - out[3] = _mm_shuffle_ps(tmp2, tmp3, 0xDD); -} - -GLM_FUNC_QUALIFIER __m128 sse_slow_det_ps(__m128 const in[4]) -{ - __m128 Fac0; - { - // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; - // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; - // valType SubFactor06 = m[1][2] * m[3][3] - m[3][2] * m[1][3]; - // valType SubFactor13 = m[1][2] * m[2][3] - m[2][2] * m[1][3]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac0 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac1; - { - // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; - // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; - // valType SubFactor07 = m[1][1] * m[3][3] - m[3][1] * m[1][3]; - // valType SubFactor14 = m[1][1] * m[2][3] - m[2][1] * m[1][3]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac1 = _mm_sub_ps(Mul00, Mul01); - } - - - __m128 Fac2; - { - // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; - // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; - // valType SubFactor08 = m[1][1] * m[3][2] - m[3][1] * m[1][2]; - // valType SubFactor15 = m[1][1] * m[2][2] - m[2][1] * m[1][2]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac2 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac3; - { - // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; - // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; - // valType SubFactor09 = m[1][0] * m[3][3] - m[3][0] * m[1][3]; - // valType SubFactor16 = m[1][0] * m[2][3] - m[2][0] * m[1][3]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac3 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac4; - { - // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; - // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; - // valType SubFactor10 = m[1][0] * m[3][2] - m[3][0] * m[1][2]; - // valType SubFactor17 = m[1][0] * m[2][2] - m[2][0] * m[1][2]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac4 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac5; - { - // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; - // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; - // valType SubFactor12 = m[1][0] * m[3][1] - m[3][0] * m[1][1]; - // valType SubFactor18 = m[1][0] * m[2][1] - m[2][0] * m[1][1]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac5 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 SignA = _mm_set_ps( 1.0f,-1.0f, 1.0f,-1.0f); - __m128 SignB = _mm_set_ps(-1.0f, 1.0f,-1.0f, 1.0f); - - // m[1][0] - // m[0][0] - // m[0][0] - // m[0][0] - __m128 Temp0 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Vec0 = _mm_shuffle_ps(Temp0, Temp0, _MM_SHUFFLE(2, 2, 2, 0)); - - // m[1][1] - // m[0][1] - // m[0][1] - // m[0][1] - __m128 Temp1 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Vec1 = _mm_shuffle_ps(Temp1, Temp1, _MM_SHUFFLE(2, 2, 2, 0)); - - // m[1][2] - // m[0][2] - // m[0][2] - // m[0][2] - __m128 Temp2 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Vec2 = _mm_shuffle_ps(Temp2, Temp2, _MM_SHUFFLE(2, 2, 2, 0)); - - // m[1][3] - // m[0][3] - // m[0][3] - // m[0][3] - __m128 Temp3 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Vec3 = _mm_shuffle_ps(Temp3, Temp3, _MM_SHUFFLE(2, 2, 2, 0)); - - // col0 - // + (Vec1[0] * Fac0[0] - Vec2[0] * Fac1[0] + Vec3[0] * Fac2[0]), - // - (Vec1[1] * Fac0[1] - Vec2[1] * Fac1[1] + Vec3[1] * Fac2[1]), - // + (Vec1[2] * Fac0[2] - Vec2[2] * Fac1[2] + Vec3[2] * Fac2[2]), - // - (Vec1[3] * Fac0[3] - Vec2[3] * Fac1[3] + Vec3[3] * Fac2[3]), - __m128 Mul00 = _mm_mul_ps(Vec1, Fac0); - __m128 Mul01 = _mm_mul_ps(Vec2, Fac1); - __m128 Mul02 = _mm_mul_ps(Vec3, Fac2); - __m128 Sub00 = _mm_sub_ps(Mul00, Mul01); - __m128 Add00 = _mm_add_ps(Sub00, Mul02); - __m128 Inv0 = _mm_mul_ps(SignB, Add00); - - // col1 - // - (Vec0[0] * Fac0[0] - Vec2[0] * Fac3[0] + Vec3[0] * Fac4[0]), - // + (Vec0[0] * Fac0[1] - Vec2[1] * Fac3[1] + Vec3[1] * Fac4[1]), - // - (Vec0[0] * Fac0[2] - Vec2[2] * Fac3[2] + Vec3[2] * Fac4[2]), - // + (Vec0[0] * Fac0[3] - Vec2[3] * Fac3[3] + Vec3[3] * Fac4[3]), - __m128 Mul03 = _mm_mul_ps(Vec0, Fac0); - __m128 Mul04 = _mm_mul_ps(Vec2, Fac3); - __m128 Mul05 = _mm_mul_ps(Vec3, Fac4); - __m128 Sub01 = _mm_sub_ps(Mul03, Mul04); - __m128 Add01 = _mm_add_ps(Sub01, Mul05); - __m128 Inv1 = _mm_mul_ps(SignA, Add01); - - // col2 - // + (Vec0[0] * Fac1[0] - Vec1[0] * Fac3[0] + Vec3[0] * Fac5[0]), - // - (Vec0[0] * Fac1[1] - Vec1[1] * Fac3[1] + Vec3[1] * Fac5[1]), - // + (Vec0[0] * Fac1[2] - Vec1[2] * Fac3[2] + Vec3[2] * Fac5[2]), - // - (Vec0[0] * Fac1[3] - Vec1[3] * Fac3[3] + Vec3[3] * Fac5[3]), - __m128 Mul06 = _mm_mul_ps(Vec0, Fac1); - __m128 Mul07 = _mm_mul_ps(Vec1, Fac3); - __m128 Mul08 = _mm_mul_ps(Vec3, Fac5); - __m128 Sub02 = _mm_sub_ps(Mul06, Mul07); - __m128 Add02 = _mm_add_ps(Sub02, Mul08); - __m128 Inv2 = _mm_mul_ps(SignB, Add02); - - // col3 - // - (Vec1[0] * Fac2[0] - Vec1[0] * Fac4[0] + Vec2[0] * Fac5[0]), - // + (Vec1[0] * Fac2[1] - Vec1[1] * Fac4[1] + Vec2[1] * Fac5[1]), - // - (Vec1[0] * Fac2[2] - Vec1[2] * Fac4[2] + Vec2[2] * Fac5[2]), - // + (Vec1[0] * Fac2[3] - Vec1[3] * Fac4[3] + Vec2[3] * Fac5[3])); - __m128 Mul09 = _mm_mul_ps(Vec0, Fac2); - __m128 Mul10 = _mm_mul_ps(Vec1, Fac4); - __m128 Mul11 = _mm_mul_ps(Vec2, Fac5); - __m128 Sub03 = _mm_sub_ps(Mul09, Mul10); - __m128 Add03 = _mm_add_ps(Sub03, Mul11); - __m128 Inv3 = _mm_mul_ps(SignA, Add03); - - __m128 Row0 = _mm_shuffle_ps(Inv0, Inv1, _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Row1 = _mm_shuffle_ps(Inv2, Inv3, _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Row2 = _mm_shuffle_ps(Row0, Row1, _MM_SHUFFLE(2, 0, 2, 0)); - - // valType Determinant = m[0][0] * Inverse[0][0] - // + m[0][1] * Inverse[1][0] - // + m[0][2] * Inverse[2][0] - // + m[0][3] * Inverse[3][0]; - __m128 Det0 = sse_dot_ps(in[0], Row2); - return Det0; -} - -GLM_FUNC_QUALIFIER __m128 sse_detd_ps -( - __m128 const m[4] -) -{ - // _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128( - - //T SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; - //T SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; - //T SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; - //T SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; - //T SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; - //T SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; - - // First 2 columns - __m128 Swp2A = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[2]), _MM_SHUFFLE(0, 1, 1, 2))); - __m128 Swp3A = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[3]), _MM_SHUFFLE(3, 2, 3, 3))); - __m128 MulA = _mm_mul_ps(Swp2A, Swp3A); - - // Second 2 columns - __m128 Swp2B = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[2]), _MM_SHUFFLE(3, 2, 3, 3))); - __m128 Swp3B = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[3]), _MM_SHUFFLE(0, 1, 1, 2))); - __m128 MulB = _mm_mul_ps(Swp2B, Swp3B); - - // Columns subtraction - __m128 SubE = _mm_sub_ps(MulA, MulB); - - // Last 2 rows - __m128 Swp2C = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[2]), _MM_SHUFFLE(0, 0, 1, 2))); - __m128 Swp3C = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[3]), _MM_SHUFFLE(1, 2, 0, 0))); - __m128 MulC = _mm_mul_ps(Swp2C, Swp3C); - __m128 SubF = _mm_sub_ps(_mm_movehl_ps(MulC, MulC), MulC); - - //tvec4 DetCof( - // + (m[1][1] * SubFactor00 - m[1][2] * SubFactor01 + m[1][3] * SubFactor02), - // - (m[1][0] * SubFactor00 - m[1][2] * SubFactor03 + m[1][3] * SubFactor04), - // + (m[1][0] * SubFactor01 - m[1][1] * SubFactor03 + m[1][3] * SubFactor05), - // - (m[1][0] * SubFactor02 - m[1][1] * SubFactor04 + m[1][2] * SubFactor05)); - - __m128 SubFacA = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(SubE), _MM_SHUFFLE(2, 1, 0, 0))); - __m128 SwpFacA = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[1]), _MM_SHUFFLE(0, 0, 0, 1))); - __m128 MulFacA = _mm_mul_ps(SwpFacA, SubFacA); - - __m128 SubTmpB = _mm_shuffle_ps(SubE, SubF, _MM_SHUFFLE(0, 0, 3, 1)); - __m128 SubFacB = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(SubTmpB), _MM_SHUFFLE(3, 1, 1, 0)));//SubF[0], SubE[3], SubE[3], SubE[1]; - __m128 SwpFacB = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[1]), _MM_SHUFFLE(1, 1, 2, 2))); - __m128 MulFacB = _mm_mul_ps(SwpFacB, SubFacB); - - __m128 SubRes = _mm_sub_ps(MulFacA, MulFacB); - - __m128 SubTmpC = _mm_shuffle_ps(SubE, SubF, _MM_SHUFFLE(1, 0, 2, 2)); - __m128 SubFacC = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(SubTmpC), _MM_SHUFFLE(3, 3, 2, 0))); - __m128 SwpFacC = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(m[1]), _MM_SHUFFLE(2, 3, 3, 3))); - __m128 MulFacC = _mm_mul_ps(SwpFacC, SubFacC); - - __m128 AddRes = _mm_add_ps(SubRes, MulFacC); - __m128 DetCof = _mm_mul_ps(AddRes, _mm_setr_ps( 1.0f,-1.0f, 1.0f,-1.0f)); - - //return m[0][0] * DetCof[0] - // + m[0][1] * DetCof[1] - // + m[0][2] * DetCof[2] - // + m[0][3] * DetCof[3]; - - return sse_dot_ps(m[0], DetCof); -} - -GLM_FUNC_QUALIFIER __m128 sse_det_ps -( - __m128 const m[4] -) -{ - // _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(add) - - //T SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; - //T SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; - //T SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; - //T SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; - //T SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; - //T SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; - - // First 2 columns - __m128 Swp2A = _mm_shuffle_ps(m[2], m[2], _MM_SHUFFLE(0, 1, 1, 2)); - __m128 Swp3A = _mm_shuffle_ps(m[3], m[3], _MM_SHUFFLE(3, 2, 3, 3)); - __m128 MulA = _mm_mul_ps(Swp2A, Swp3A); - - // Second 2 columns - __m128 Swp2B = _mm_shuffle_ps(m[2], m[2], _MM_SHUFFLE(3, 2, 3, 3)); - __m128 Swp3B = _mm_shuffle_ps(m[3], m[3], _MM_SHUFFLE(0, 1, 1, 2)); - __m128 MulB = _mm_mul_ps(Swp2B, Swp3B); - - // Columns subtraction - __m128 SubE = _mm_sub_ps(MulA, MulB); - - // Last 2 rows - __m128 Swp2C = _mm_shuffle_ps(m[2], m[2], _MM_SHUFFLE(0, 0, 1, 2)); - __m128 Swp3C = _mm_shuffle_ps(m[3], m[3], _MM_SHUFFLE(1, 2, 0, 0)); - __m128 MulC = _mm_mul_ps(Swp2C, Swp3C); - __m128 SubF = _mm_sub_ps(_mm_movehl_ps(MulC, MulC), MulC); - - //tvec4 DetCof( - // + (m[1][1] * SubFactor00 - m[1][2] * SubFactor01 + m[1][3] * SubFactor02), - // - (m[1][0] * SubFactor00 - m[1][2] * SubFactor03 + m[1][3] * SubFactor04), - // + (m[1][0] * SubFactor01 - m[1][1] * SubFactor03 + m[1][3] * SubFactor05), - // - (m[1][0] * SubFactor02 - m[1][1] * SubFactor04 + m[1][2] * SubFactor05)); - - __m128 SubFacA = _mm_shuffle_ps(SubE, SubE, _MM_SHUFFLE(2, 1, 0, 0)); - __m128 SwpFacA = _mm_shuffle_ps(m[1], m[1], _MM_SHUFFLE(0, 0, 0, 1)); - __m128 MulFacA = _mm_mul_ps(SwpFacA, SubFacA); - - __m128 SubTmpB = _mm_shuffle_ps(SubE, SubF, _MM_SHUFFLE(0, 0, 3, 1)); - __m128 SubFacB = _mm_shuffle_ps(SubTmpB, SubTmpB, _MM_SHUFFLE(3, 1, 1, 0));//SubF[0], SubE[3], SubE[3], SubE[1]; - __m128 SwpFacB = _mm_shuffle_ps(m[1], m[1], _MM_SHUFFLE(1, 1, 2, 2)); - __m128 MulFacB = _mm_mul_ps(SwpFacB, SubFacB); - - __m128 SubRes = _mm_sub_ps(MulFacA, MulFacB); - - __m128 SubTmpC = _mm_shuffle_ps(SubE, SubF, _MM_SHUFFLE(1, 0, 2, 2)); - __m128 SubFacC = _mm_shuffle_ps(SubTmpC, SubTmpC, _MM_SHUFFLE(3, 3, 2, 0)); - __m128 SwpFacC = _mm_shuffle_ps(m[1], m[1], _MM_SHUFFLE(2, 3, 3, 3)); - __m128 MulFacC = _mm_mul_ps(SwpFacC, SubFacC); - - __m128 AddRes = _mm_add_ps(SubRes, MulFacC); - __m128 DetCof = _mm_mul_ps(AddRes, _mm_setr_ps( 1.0f,-1.0f, 1.0f,-1.0f)); - - //return m[0][0] * DetCof[0] - // + m[0][1] * DetCof[1] - // + m[0][2] * DetCof[2] - // + m[0][3] * DetCof[3]; - - return sse_dot_ps(m[0], DetCof); -} - -GLM_FUNC_QUALIFIER void sse_inverse_ps(__m128 const in[4], __m128 out[4]) -{ - __m128 Fac0; - { - // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; - // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; - // valType SubFactor06 = m[1][2] * m[3][3] - m[3][2] * m[1][3]; - // valType SubFactor13 = m[1][2] * m[2][3] - m[2][2] * m[1][3]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac0 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac1; - { - // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; - // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; - // valType SubFactor07 = m[1][1] * m[3][3] - m[3][1] * m[1][3]; - // valType SubFactor14 = m[1][1] * m[2][3] - m[2][1] * m[1][3]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac1 = _mm_sub_ps(Mul00, Mul01); - } - - - __m128 Fac2; - { - // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; - // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; - // valType SubFactor08 = m[1][1] * m[3][2] - m[3][1] * m[1][2]; - // valType SubFactor15 = m[1][1] * m[2][2] - m[2][1] * m[1][2]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac2 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac3; - { - // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; - // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; - // valType SubFactor09 = m[1][0] * m[3][3] - m[3][0] * m[1][3]; - // valType SubFactor16 = m[1][0] * m[2][3] - m[2][0] * m[1][3]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac3 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac4; - { - // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; - // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; - // valType SubFactor10 = m[1][0] * m[3][2] - m[3][0] * m[1][2]; - // valType SubFactor17 = m[1][0] * m[2][2] - m[2][0] * m[1][2]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac4 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac5; - { - // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; - // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; - // valType SubFactor12 = m[1][0] * m[3][1] - m[3][0] * m[1][1]; - // valType SubFactor18 = m[1][0] * m[2][1] - m[2][0] * m[1][1]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac5 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 SignA = _mm_set_ps( 1.0f,-1.0f, 1.0f,-1.0f); - __m128 SignB = _mm_set_ps(-1.0f, 1.0f,-1.0f, 1.0f); - - // m[1][0] - // m[0][0] - // m[0][0] - // m[0][0] - __m128 Temp0 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Vec0 = _mm_shuffle_ps(Temp0, Temp0, _MM_SHUFFLE(2, 2, 2, 0)); - - // m[1][1] - // m[0][1] - // m[0][1] - // m[0][1] - __m128 Temp1 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Vec1 = _mm_shuffle_ps(Temp1, Temp1, _MM_SHUFFLE(2, 2, 2, 0)); - - // m[1][2] - // m[0][2] - // m[0][2] - // m[0][2] - __m128 Temp2 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Vec2 = _mm_shuffle_ps(Temp2, Temp2, _MM_SHUFFLE(2, 2, 2, 0)); - - // m[1][3] - // m[0][3] - // m[0][3] - // m[0][3] - __m128 Temp3 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Vec3 = _mm_shuffle_ps(Temp3, Temp3, _MM_SHUFFLE(2, 2, 2, 0)); - - // col0 - // + (Vec1[0] * Fac0[0] - Vec2[0] * Fac1[0] + Vec3[0] * Fac2[0]), - // - (Vec1[1] * Fac0[1] - Vec2[1] * Fac1[1] + Vec3[1] * Fac2[1]), - // + (Vec1[2] * Fac0[2] - Vec2[2] * Fac1[2] + Vec3[2] * Fac2[2]), - // - (Vec1[3] * Fac0[3] - Vec2[3] * Fac1[3] + Vec3[3] * Fac2[3]), - __m128 Mul00 = _mm_mul_ps(Vec1, Fac0); - __m128 Mul01 = _mm_mul_ps(Vec2, Fac1); - __m128 Mul02 = _mm_mul_ps(Vec3, Fac2); - __m128 Sub00 = _mm_sub_ps(Mul00, Mul01); - __m128 Add00 = _mm_add_ps(Sub00, Mul02); - __m128 Inv0 = _mm_mul_ps(SignB, Add00); - - // col1 - // - (Vec0[0] * Fac0[0] - Vec2[0] * Fac3[0] + Vec3[0] * Fac4[0]), - // + (Vec0[0] * Fac0[1] - Vec2[1] * Fac3[1] + Vec3[1] * Fac4[1]), - // - (Vec0[0] * Fac0[2] - Vec2[2] * Fac3[2] + Vec3[2] * Fac4[2]), - // + (Vec0[0] * Fac0[3] - Vec2[3] * Fac3[3] + Vec3[3] * Fac4[3]), - __m128 Mul03 = _mm_mul_ps(Vec0, Fac0); - __m128 Mul04 = _mm_mul_ps(Vec2, Fac3); - __m128 Mul05 = _mm_mul_ps(Vec3, Fac4); - __m128 Sub01 = _mm_sub_ps(Mul03, Mul04); - __m128 Add01 = _mm_add_ps(Sub01, Mul05); - __m128 Inv1 = _mm_mul_ps(SignA, Add01); - - // col2 - // + (Vec0[0] * Fac1[0] - Vec1[0] * Fac3[0] + Vec3[0] * Fac5[0]), - // - (Vec0[0] * Fac1[1] - Vec1[1] * Fac3[1] + Vec3[1] * Fac5[1]), - // + (Vec0[0] * Fac1[2] - Vec1[2] * Fac3[2] + Vec3[2] * Fac5[2]), - // - (Vec0[0] * Fac1[3] - Vec1[3] * Fac3[3] + Vec3[3] * Fac5[3]), - __m128 Mul06 = _mm_mul_ps(Vec0, Fac1); - __m128 Mul07 = _mm_mul_ps(Vec1, Fac3); - __m128 Mul08 = _mm_mul_ps(Vec3, Fac5); - __m128 Sub02 = _mm_sub_ps(Mul06, Mul07); - __m128 Add02 = _mm_add_ps(Sub02, Mul08); - __m128 Inv2 = _mm_mul_ps(SignB, Add02); - - // col3 - // - (Vec1[0] * Fac2[0] - Vec1[0] * Fac4[0] + Vec2[0] * Fac5[0]), - // + (Vec1[0] * Fac2[1] - Vec1[1] * Fac4[1] + Vec2[1] * Fac5[1]), - // - (Vec1[0] * Fac2[2] - Vec1[2] * Fac4[2] + Vec2[2] * Fac5[2]), - // + (Vec1[0] * Fac2[3] - Vec1[3] * Fac4[3] + Vec2[3] * Fac5[3])); - __m128 Mul09 = _mm_mul_ps(Vec0, Fac2); - __m128 Mul10 = _mm_mul_ps(Vec1, Fac4); - __m128 Mul11 = _mm_mul_ps(Vec2, Fac5); - __m128 Sub03 = _mm_sub_ps(Mul09, Mul10); - __m128 Add03 = _mm_add_ps(Sub03, Mul11); - __m128 Inv3 = _mm_mul_ps(SignA, Add03); - - __m128 Row0 = _mm_shuffle_ps(Inv0, Inv1, _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Row1 = _mm_shuffle_ps(Inv2, Inv3, _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Row2 = _mm_shuffle_ps(Row0, Row1, _MM_SHUFFLE(2, 0, 2, 0)); - - // valType Determinant = m[0][0] * Inverse[0][0] - // + m[0][1] * Inverse[1][0] - // + m[0][2] * Inverse[2][0] - // + m[0][3] * Inverse[3][0]; - __m128 Det0 = sse_dot_ps(in[0], Row2); - __m128 Rcp0 = _mm_div_ps(one, Det0); - //__m128 Rcp0 = _mm_rcp_ps(Det0); - - // Inverse /= Determinant; - out[0] = _mm_mul_ps(Inv0, Rcp0); - out[1] = _mm_mul_ps(Inv1, Rcp0); - out[2] = _mm_mul_ps(Inv2, Rcp0); - out[3] = _mm_mul_ps(Inv3, Rcp0); -} - -GLM_FUNC_QUALIFIER void sse_inverse_fast_ps(__m128 const in[4], __m128 out[4]) -{ - __m128 Fac0; - { - // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; - // valType SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; - // valType SubFactor06 = m[1][2] * m[3][3] - m[3][2] * m[1][3]; - // valType SubFactor13 = m[1][2] * m[2][3] - m[2][2] * m[1][3]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac0 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac1; - { - // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; - // valType SubFactor01 = m[2][1] * m[3][3] - m[3][1] * m[2][3]; - // valType SubFactor07 = m[1][1] * m[3][3] - m[3][1] * m[1][3]; - // valType SubFactor14 = m[1][1] * m[2][3] - m[2][1] * m[1][3]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac1 = _mm_sub_ps(Mul00, Mul01); - } - - - __m128 Fac2; - { - // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; - // valType SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2]; - // valType SubFactor08 = m[1][1] * m[3][2] - m[3][1] * m[1][2]; - // valType SubFactor15 = m[1][1] * m[2][2] - m[2][1] * m[1][2]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac2 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac3; - { - // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; - // valType SubFactor03 = m[2][0] * m[3][3] - m[3][0] * m[2][3]; - // valType SubFactor09 = m[1][0] * m[3][3] - m[3][0] * m[1][3]; - // valType SubFactor16 = m[1][0] * m[2][3] - m[2][0] * m[1][3]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(3, 3, 3, 3)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac3 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac4; - { - // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; - // valType SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; - // valType SubFactor10 = m[1][0] * m[3][2] - m[3][0] * m[1][2]; - // valType SubFactor17 = m[1][0] * m[2][2] - m[2][0] * m[1][2]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(2, 2, 2, 2)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac4 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 Fac5; - { - // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; - // valType SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1]; - // valType SubFactor12 = m[1][0] * m[3][1] - m[3][0] * m[1][1]; - // valType SubFactor18 = m[1][0] * m[2][1] - m[2][0] * m[1][1]; - - __m128 Swp0a = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Swp0b = _mm_shuffle_ps(in[3], in[2], _MM_SHUFFLE(0, 0, 0, 0)); - - __m128 Swp00 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Swp01 = _mm_shuffle_ps(Swp0a, Swp0a, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp02 = _mm_shuffle_ps(Swp0b, Swp0b, _MM_SHUFFLE(2, 0, 0, 0)); - __m128 Swp03 = _mm_shuffle_ps(in[2], in[1], _MM_SHUFFLE(1, 1, 1, 1)); - - __m128 Mul00 = _mm_mul_ps(Swp00, Swp01); - __m128 Mul01 = _mm_mul_ps(Swp02, Swp03); - Fac5 = _mm_sub_ps(Mul00, Mul01); - } - - __m128 SignA = _mm_set_ps( 1.0f,-1.0f, 1.0f,-1.0f); - __m128 SignB = _mm_set_ps(-1.0f, 1.0f,-1.0f, 1.0f); - - // m[1][0] - // m[0][0] - // m[0][0] - // m[0][0] - __m128 Temp0 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Vec0 = _mm_shuffle_ps(Temp0, Temp0, _MM_SHUFFLE(2, 2, 2, 0)); - - // m[1][1] - // m[0][1] - // m[0][1] - // m[0][1] - __m128 Temp1 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(1, 1, 1, 1)); - __m128 Vec1 = _mm_shuffle_ps(Temp1, Temp1, _MM_SHUFFLE(2, 2, 2, 0)); - - // m[1][2] - // m[0][2] - // m[0][2] - // m[0][2] - __m128 Temp2 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(2, 2, 2, 2)); - __m128 Vec2 = _mm_shuffle_ps(Temp2, Temp2, _MM_SHUFFLE(2, 2, 2, 0)); - - // m[1][3] - // m[0][3] - // m[0][3] - // m[0][3] - __m128 Temp3 = _mm_shuffle_ps(in[1], in[0], _MM_SHUFFLE(3, 3, 3, 3)); - __m128 Vec3 = _mm_shuffle_ps(Temp3, Temp3, _MM_SHUFFLE(2, 2, 2, 0)); - - // col0 - // + (Vec1[0] * Fac0[0] - Vec2[0] * Fac1[0] + Vec3[0] * Fac2[0]), - // - (Vec1[1] * Fac0[1] - Vec2[1] * Fac1[1] + Vec3[1] * Fac2[1]), - // + (Vec1[2] * Fac0[2] - Vec2[2] * Fac1[2] + Vec3[2] * Fac2[2]), - // - (Vec1[3] * Fac0[3] - Vec2[3] * Fac1[3] + Vec3[3] * Fac2[3]), - __m128 Mul00 = _mm_mul_ps(Vec1, Fac0); - __m128 Mul01 = _mm_mul_ps(Vec2, Fac1); - __m128 Mul02 = _mm_mul_ps(Vec3, Fac2); - __m128 Sub00 = _mm_sub_ps(Mul00, Mul01); - __m128 Add00 = _mm_add_ps(Sub00, Mul02); - __m128 Inv0 = _mm_mul_ps(SignB, Add00); - - // col1 - // - (Vec0[0] * Fac0[0] - Vec2[0] * Fac3[0] + Vec3[0] * Fac4[0]), - // + (Vec0[0] * Fac0[1] - Vec2[1] * Fac3[1] + Vec3[1] * Fac4[1]), - // - (Vec0[0] * Fac0[2] - Vec2[2] * Fac3[2] + Vec3[2] * Fac4[2]), - // + (Vec0[0] * Fac0[3] - Vec2[3] * Fac3[3] + Vec3[3] * Fac4[3]), - __m128 Mul03 = _mm_mul_ps(Vec0, Fac0); - __m128 Mul04 = _mm_mul_ps(Vec2, Fac3); - __m128 Mul05 = _mm_mul_ps(Vec3, Fac4); - __m128 Sub01 = _mm_sub_ps(Mul03, Mul04); - __m128 Add01 = _mm_add_ps(Sub01, Mul05); - __m128 Inv1 = _mm_mul_ps(SignA, Add01); - - // col2 - // + (Vec0[0] * Fac1[0] - Vec1[0] * Fac3[0] + Vec3[0] * Fac5[0]), - // - (Vec0[0] * Fac1[1] - Vec1[1] * Fac3[1] + Vec3[1] * Fac5[1]), - // + (Vec0[0] * Fac1[2] - Vec1[2] * Fac3[2] + Vec3[2] * Fac5[2]), - // - (Vec0[0] * Fac1[3] - Vec1[3] * Fac3[3] + Vec3[3] * Fac5[3]), - __m128 Mul06 = _mm_mul_ps(Vec0, Fac1); - __m128 Mul07 = _mm_mul_ps(Vec1, Fac3); - __m128 Mul08 = _mm_mul_ps(Vec3, Fac5); - __m128 Sub02 = _mm_sub_ps(Mul06, Mul07); - __m128 Add02 = _mm_add_ps(Sub02, Mul08); - __m128 Inv2 = _mm_mul_ps(SignB, Add02); - - // col3 - // - (Vec1[0] * Fac2[0] - Vec1[0] * Fac4[0] + Vec2[0] * Fac5[0]), - // + (Vec1[0] * Fac2[1] - Vec1[1] * Fac4[1] + Vec2[1] * Fac5[1]), - // - (Vec1[0] * Fac2[2] - Vec1[2] * Fac4[2] + Vec2[2] * Fac5[2]), - // + (Vec1[0] * Fac2[3] - Vec1[3] * Fac4[3] + Vec2[3] * Fac5[3])); - __m128 Mul09 = _mm_mul_ps(Vec0, Fac2); - __m128 Mul10 = _mm_mul_ps(Vec1, Fac4); - __m128 Mul11 = _mm_mul_ps(Vec2, Fac5); - __m128 Sub03 = _mm_sub_ps(Mul09, Mul10); - __m128 Add03 = _mm_add_ps(Sub03, Mul11); - __m128 Inv3 = _mm_mul_ps(SignA, Add03); - - __m128 Row0 = _mm_shuffle_ps(Inv0, Inv1, _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Row1 = _mm_shuffle_ps(Inv2, Inv3, _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Row2 = _mm_shuffle_ps(Row0, Row1, _MM_SHUFFLE(2, 0, 2, 0)); - - // valType Determinant = m[0][0] * Inverse[0][0] - // + m[0][1] * Inverse[1][0] - // + m[0][2] * Inverse[2][0] - // + m[0][3] * Inverse[3][0]; - __m128 Det0 = sse_dot_ps(in[0], Row2); - __m128 Rcp0 = _mm_rcp_ps(Det0); - //__m128 Rcp0 = _mm_div_ps(one, Det0); - // Inverse /= Determinant; - out[0] = _mm_mul_ps(Inv0, Rcp0); - out[1] = _mm_mul_ps(Inv1, Rcp0); - out[2] = _mm_mul_ps(Inv2, Rcp0); - out[3] = _mm_mul_ps(Inv3, Rcp0); -} -/* -GLM_FUNC_QUALIFIER void sse_rotate_ps(__m128 const in[4], float Angle, float const v[3], __m128 out[4]) -{ - float a = glm::radians(Angle); - float c = cos(a); - float s = sin(a); - - glm::vec4 AxisA(v[0], v[1], v[2], float(0)); - __m128 AxisB = _mm_set_ps(AxisA.w, AxisA.z, AxisA.y, AxisA.x); - __m128 AxisC = detail::sse_nrm_ps(AxisB); - - __m128 Cos0 = _mm_set_ss(c); - __m128 CosA = _mm_shuffle_ps(Cos0, Cos0, _MM_SHUFFLE(0, 0, 0, 0)); - __m128 Sin0 = _mm_set_ss(s); - __m128 SinA = _mm_shuffle_ps(Sin0, Sin0, _MM_SHUFFLE(0, 0, 0, 0)); - - // tvec3 temp = (valType(1) - c) * axis; - __m128 Temp0 = _mm_sub_ps(one, CosA); - __m128 Temp1 = _mm_mul_ps(Temp0, AxisC); - - //Rotate[0][0] = c + temp[0] * axis[0]; - //Rotate[0][1] = 0 + temp[0] * axis[1] + s * axis[2]; - //Rotate[0][2] = 0 + temp[0] * axis[2] - s * axis[1]; - __m128 Axis0 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(0, 0, 0, 0)); - __m128 TmpA0 = _mm_mul_ps(Axis0, AxisC); - __m128 CosA0 = _mm_shuffle_ps(Cos0, Cos0, _MM_SHUFFLE(1, 1, 1, 0)); - __m128 TmpA1 = _mm_add_ps(CosA0, TmpA0); - __m128 SinA0 = SinA;//_mm_set_ps(0.0f, s, -s, 0.0f); - __m128 TmpA2 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(3, 1, 2, 3)); - __m128 TmpA3 = _mm_mul_ps(SinA0, TmpA2); - __m128 TmpA4 = _mm_add_ps(TmpA1, TmpA3); - - //Rotate[1][0] = 0 + temp[1] * axis[0] - s * axis[2]; - //Rotate[1][1] = c + temp[1] * axis[1]; - //Rotate[1][2] = 0 + temp[1] * axis[2] + s * axis[0]; - __m128 Axis1 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(1, 1, 1, 1)); - __m128 TmpB0 = _mm_mul_ps(Axis1, AxisC); - __m128 CosA1 = _mm_shuffle_ps(Cos0, Cos0, _MM_SHUFFLE(1, 1, 0, 1)); - __m128 TmpB1 = _mm_add_ps(CosA1, TmpB0); - __m128 SinB0 = SinA;//_mm_set_ps(-s, 0.0f, s, 0.0f); - __m128 TmpB2 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(3, 0, 3, 2)); - __m128 TmpB3 = _mm_mul_ps(SinA0, TmpB2); - __m128 TmpB4 = _mm_add_ps(TmpB1, TmpB3); - - //Rotate[2][0] = 0 + temp[2] * axis[0] + s * axis[1]; - //Rotate[2][1] = 0 + temp[2] * axis[1] - s * axis[0]; - //Rotate[2][2] = c + temp[2] * axis[2]; - __m128 Axis2 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(2, 2, 2, 2)); - __m128 TmpC0 = _mm_mul_ps(Axis2, AxisC); - __m128 CosA2 = _mm_shuffle_ps(Cos0, Cos0, _MM_SHUFFLE(1, 0, 1, 1)); - __m128 TmpC1 = _mm_add_ps(CosA2, TmpC0); - __m128 SinC0 = SinA;//_mm_set_ps(s, -s, 0.0f, 0.0f); - __m128 TmpC2 = _mm_shuffle_ps(AxisC, AxisC, _MM_SHUFFLE(3, 3, 0, 1)); - __m128 TmpC3 = _mm_mul_ps(SinA0, TmpC2); - __m128 TmpC4 = _mm_add_ps(TmpC1, TmpC3); - - __m128 Result[4]; - Result[0] = TmpA4; - Result[1] = TmpB4; - Result[2] = TmpC4; - Result[3] = _mm_set_ps(1, 0, 0, 0); - - //tmat4x4 Result(uninitialize); - //Result[0] = m[0] * Rotate[0][0] + m[1] * Rotate[0][1] + m[2] * Rotate[0][2]; - //Result[1] = m[0] * Rotate[1][0] + m[1] * Rotate[1][1] + m[2] * Rotate[1][2]; - //Result[2] = m[0] * Rotate[2][0] + m[1] * Rotate[2][1] + m[2] * Rotate[2][2]; - //Result[3] = m[3]; - //return Result; - sse_mul_ps(in, Result, out); -} -*/ -GLM_FUNC_QUALIFIER void sse_outer_ps(__m128 const & c, __m128 const & r, __m128 out[4]) -{ - out[0] = _mm_mul_ps(c, _mm_shuffle_ps(r, r, _MM_SHUFFLE(0, 0, 0, 0))); - out[1] = _mm_mul_ps(c, _mm_shuffle_ps(r, r, _MM_SHUFFLE(1, 1, 1, 1))); - out[2] = _mm_mul_ps(c, _mm_shuffle_ps(r, r, _MM_SHUFFLE(2, 2, 2, 2))); - out[3] = _mm_mul_ps(c, _mm_shuffle_ps(r, r, _MM_SHUFFLE(3, 3, 3, 3))); -} - -}//namespace detail -}//namespace glm diff --git a/glm/detail/intrinsic_trigonometric.hpp b/glm/detail/intrinsic_trigonometric.hpp deleted file mode 100644 index bf2c9e22..00000000 --- a/glm/detail/intrinsic_trigonometric.hpp +++ /dev/null @@ -1,46 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_trigonometric.hpp -/// @date 2009-06-09 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -#pragma once - -#include "setup.hpp" - -#if(!(GLM_ARCH & GLM_ARCH_SSE2)) -# error "SSE2 instructions not supported or enabled" -#else - -namespace glm{ -namespace detail -{ - -}//namespace detail -}//namespace glm - -#include "intrinsic_trigonometric.inl" - -#endif//GLM_ARCH diff --git a/glm/detail/intrinsic_trigonometric.inl b/glm/detail/intrinsic_trigonometric.inl deleted file mode 100644 index 5941ca31..00000000 --- a/glm/detail/intrinsic_trigonometric.inl +++ /dev/null @@ -1,27 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_trigonometric.inl -/// @date 2011-06-15 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// diff --git a/glm/detail/intrinsic_vector_relational.hpp b/glm/detail/intrinsic_vector_relational.hpp deleted file mode 100644 index fe7e902b..00000000 --- a/glm/detail/intrinsic_vector_relational.hpp +++ /dev/null @@ -1,46 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_vector_relational.hpp -/// @date 2009-06-09 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -#pragma once - -#include "setup.hpp" - -#if(!(GLM_ARCH & GLM_ARCH_SSE2)) -# error "SSE2 instructions not supported or enabled" -#else - -namespace glm{ -namespace detail -{ - -}//namespace detail -}//namespace glm - -#include "intrinsic_vector_relational.inl" - -#endif//GLM_ARCH diff --git a/glm/detail/intrinsic_vector_relational.inl b/glm/detail/intrinsic_vector_relational.inl deleted file mode 100644 index e7445edc..00000000 --- a/glm/detail/intrinsic_vector_relational.inl +++ /dev/null @@ -1,366 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/intrinsic_vector_relational.inl -/// @date 2009-06-09 / 2011-06-15 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// -// -//// lessThan -//template -//GLM_FUNC_QUALIFIER typename tvec2::bool_type lessThan -//( -// tvec2 const & x, -// tvec2 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec2::bool_type(x.x < y.x, x.y < y.y); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec3::bool_type lessThan -//( -// tvec3 const & x, -// tvec3 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec3::bool_type(x.x < y.x, x.y < y.y, x.z < y.z); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec4::bool_type lessThan -//( -// tvec4 const & x, -// tvec4 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec4::bool_type(x.x < y.x, x.y < y.y, x.z < y.z, x.w < y.w); -//} -// -//// lessThanEqual -//template -//GLM_FUNC_QUALIFIER typename tvec2::bool_type lessThanEqual -//( -// tvec2 const & x, -// tvec2 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec2::bool_type(x.x <= y.x, x.y <= y.y); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec3::bool_type lessThanEqual -//( -// tvec3 const & x, -// tvec3 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec3::bool_type(x.x <= y.x, x.y <= y.y, x.z <= y.z); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec4::bool_type lessThanEqual -//( -// tvec4 const & x, -// tvec4 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec4::bool_type(x.x <= y.x, x.y <= y.y, x.z <= y.z, x.w <= y.w); -//} -// -//// greaterThan -//template -//GLM_FUNC_QUALIFIER typename tvec2::bool_type greaterThan -//( -// tvec2 const & x, -// tvec2 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec2::bool_type(x.x > y.x, x.y > y.y); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec3::bool_type greaterThan -//( -// tvec3 const & x, -// tvec3 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec3::bool_type(x.x > y.x, x.y > y.y, x.z > y.z); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec4::bool_type greaterThan -//( -// tvec4 const & x, -// tvec4 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec4::bool_type(x.x > y.x, x.y > y.y, x.z > y.z, x.w > y.w); -//} -// -//// greaterThanEqual -//template -//GLM_FUNC_QUALIFIER typename tvec2::bool_type greaterThanEqual -//( -// tvec2 const & x, -// tvec2 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec2::bool_type(x.x >= y.x, x.y >= y.y); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec3::bool_type greaterThanEqual -//( -// tvec3 const & x, -// tvec3 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec3::bool_type(x.x >= y.x, x.y >= y.y, x.z >= y.z); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec4::bool_type greaterThanEqual -//( -// tvec4 const & x, -// tvec4 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint); -// -// return typename tvec4::bool_type(x.x >= y.x, x.y >= y.y, x.z >= y.z, x.w >= y.w); -//} -// -//// equal -//template -//GLM_FUNC_QUALIFIER typename tvec2::bool_type equal -//( -// tvec2 const & x, -// tvec2 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint || -// detail::type::is_bool); -// -// return typename tvec2::bool_type(x.x == y.x, x.y == y.y); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec3::bool_type equal -//( -// tvec3 const & x, -// tvec3 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint || -// detail::type::is_bool); -// -// return typename tvec3::bool_type(x.x == y.x, x.y == y.y, x.z == y.z); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec4::bool_type equal -//( -// tvec4 const & x, -// tvec4 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint || -// detail::type::is_bool); -// -// return typename tvec4::bool_type(x.x == y.x, x.y == y.y, x.z == y.z, x.w == y.w); -//} -// -//// notEqual -//template -//GLM_FUNC_QUALIFIER typename tvec2::bool_type notEqual -//( -// tvec2 const & x, -// tvec2 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint || -// detail::type::is_bool); -// -// return typename tvec2::bool_type(x.x != y.x, x.y != y.y); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec3::bool_type notEqual -//( -// tvec3 const & x, -// tvec3 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint || -// detail::type::is_bool); -// -// return typename tvec3::bool_type(x.x != y.x, x.y != y.y, x.z != y.z); -//} -// -//template -//GLM_FUNC_QUALIFIER typename tvec4::bool_type notEqual -//( -// tvec4 const & x, -// tvec4 const & y -//) -//{ -// GLM_STATIC_ASSERT( -// detail::type::is_float || -// detail::type::is_int || -// detail::type::is_uint || -// detail::type::is_bool); -// -// return typename tvec4::bool_type(x.x != y.x, x.y != y.y, x.z != y.z, x.w != y.w); -//} -// -//// any -//GLM_FUNC_QUALIFIER bool any(tvec2 const & x) -//{ -// return x.x || x.y; -//} -// -//GLM_FUNC_QUALIFIER bool any(tvec3 const & x) -//{ -// return x.x || x.y || x.z; -//} -// -//GLM_FUNC_QUALIFIER bool any(tvec4 const & x) -//{ -// return x.x || x.y || x.z || x.w; -//} -// -//// all -//GLM_FUNC_QUALIFIER bool all(const tvec2& x) -//{ -// return x.x && x.y; -//} -// -//GLM_FUNC_QUALIFIER bool all(const tvec3& x) -//{ -// return x.x && x.y && x.z; -//} -// -//GLM_FUNC_QUALIFIER bool all(const tvec4& x) -//{ -// return x.x && x.y && x.z && x.w; -//} -// -//// not -//GLM_FUNC_QUALIFIER tvec2::bool_type not_ -//( -// tvec2 const & v -//) -//{ -// return tvec2::bool_type(!v.x, !v.y); -//} -// -//GLM_FUNC_QUALIFIER tvec3::bool_type not_ -//( -// tvec3 const & v -//) -//{ -// return tvec3::bool_type(!v.x, !v.y, !v.z); -//} -// -//GLM_FUNC_QUALIFIER tvec4::bool_type not_ -//( -// tvec4 const & v -//) -//{ -// return tvec4::bool_type(!v.x, !v.y, !v.z, !v.w); -//} \ No newline at end of file diff --git a/glm/detail/setup.hpp b/glm/detail/setup.hpp index 3103d32d..a98b207d 100644 --- a/glm/detail/setup.hpp +++ b/glm/detail/setup.hpp @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/detail/setup.hpp -/// @date 2006-11-13 / 2014-10-05 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/detail/type_vec4_avx.inl b/glm/detail/type_vec4_avx.inl deleted file mode 100644 index f2a44398..00000000 --- a/glm/detail/type_vec4_avx.inl +++ /dev/null @@ -1,41 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/type_tvec4_avx.inl -/// @date 2014-12-01 / 2014-12-01 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -namespace glm{ -namespace detail -{ - -}//namespace detail - - - -}//namespace glm diff --git a/glm/detail/type_vec4_avx2.inl b/glm/detail/type_vec4_avx2.inl deleted file mode 100644 index fda75e51..00000000 --- a/glm/detail/type_vec4_avx2.inl +++ /dev/null @@ -1,41 +0,0 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @ref core -/// @file glm/detail/type_tvec4_avx2.inl -/// @date 2014-12-01 / 2014-12-01 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// - -namespace glm{ -namespace detail -{ - -}//namespace detail - - - -}//namespace glm diff --git a/glm/detail/type_vec4_sse2.inl b/glm/detail/type_vec4_simd.inl similarity index 100% rename from glm/detail/type_vec4_sse2.inl rename to glm/detail/type_vec4_simd.inl diff --git a/glm/ext.hpp b/glm/ext.hpp index dc8f9d79..59476b0a 100644 --- a/glm/ext.hpp +++ b/glm/ext.hpp @@ -1,32 +1,4 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// -/// @file glm/glm.hpp -/// @date 2009-05-01 / 2011-05-16 -/// @author Christophe Riccio +/// @file glm/ext.hpp /// /// @ref core (Dependence) /// @@ -52,7 +24,6 @@ /// Even if it's highly unrecommended, it's possible to include all the extensions /// at once by including . Otherwise, each extension needs to be /// included a specific file. -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/gtc/epsilon.hpp b/glm/gtc/epsilon.hpp index b69e1a03..a76aaaf7 100644 --- a/glm/gtc/epsilon.hpp +++ b/glm/gtc/epsilon.hpp @@ -1,33 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref gtc_epsilon /// @file glm/gtc/epsilon.hpp -/// @date 2012-04-07 / 2012-04-07 -/// @author Christophe Riccio /// /// @see core (dependence) /// @see gtc_half_float (dependence) @@ -39,7 +11,6 @@ /// @brief Comparison functions for a user defined epsilon values. /// /// need to be included to use these functionalities. -/////////////////////////////////////////////////////////////////////////////////// #pragma once diff --git a/glm/vec4.hpp b/glm/vec4.hpp index 91095e8b..ad66f5ef 100644 --- a/glm/vec4.hpp +++ b/glm/vec4.hpp @@ -1,34 +1,5 @@ -/////////////////////////////////////////////////////////////////////////////////// -/// OpenGL Mathematics (glm.g-truc.net) -/// -/// Copyright (c) 2005 - 2015 G-Truc Creation (www.g-truc.net) -/// Permission is hereby granted, free of charge, to any person obtaining a copy -/// of this software and associated documentation files (the "Software"), to deal -/// in the Software without restriction, including without limitation the rights -/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell -/// copies of the Software, and to permit persons to whom the Software is -/// furnished to do so, subject to the following conditions: -/// -/// The above copyright notice and this permission notice shall be included in -/// all copies or substantial portions of the Software. -/// -/// Restrictions: -/// By making use of the Software for military purposes, you choose to make -/// a Bunny unhappy. -/// -/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN -/// THE SOFTWARE. -/// /// @ref core /// @file glm/vec4.hpp -/// @date 2013-12-24 / 2013-12-24 -/// @author Christophe Riccio -/////////////////////////////////////////////////////////////////////////////////// #pragma once