diff options
author | Dave Parks <davep@lindenlab.com> | 2011-11-08 17:46:43 -0600 |
---|---|---|
committer | Dave Parks <davep@lindenlab.com> | 2011-11-08 17:46:43 -0600 |
commit | db7924681d629f48f939e7d4cc55d588aa93c344 (patch) | |
tree | b6ba99c462b1574bcc366db1807ba0b593c66e17 /indra/llmath | |
parent | 0d568ce9f0522eaa21a12fa5fae31f80c59bd947 (diff) |
SH-24114 Remove unused vectorization code (outdated/unused with SSE2 requirement and llvetor4a et al)
Diffstat (limited to 'indra/llmath')
-rw-r--r-- | indra/llmath/CMakeLists.txt | 4 | ||||
-rw-r--r-- | indra/llmath/llv4math.h | 141 | ||||
-rw-r--r-- | indra/llmath/llv4matrix3.h | 220 | ||||
-rw-r--r-- | indra/llmath/llv4matrix4.h | 249 | ||||
-rw-r--r-- | indra/llmath/llv4vector3.h | 80 |
5 files changed, 0 insertions, 694 deletions
diff --git a/indra/llmath/CMakeLists.txt b/indra/llmath/CMakeLists.txt index cd100cdf9f..b5e59c1ca3 100644 --- a/indra/llmath/CMakeLists.txt +++ b/indra/llmath/CMakeLists.txt @@ -75,10 +75,6 @@ set(llmath_HEADER_FILES llvector4a.h llvector4a.inl llvector4logical.h - llv4math.h - llv4matrix3.h - llv4matrix4.h - llv4vector3.h llvolume.h llvolumemgr.h llvolumeoctree.h diff --git a/indra/llmath/llv4math.h b/indra/llmath/llv4math.h deleted file mode 100644 index 5f403ba526..0000000000 --- a/indra/llmath/llv4math.h +++ /dev/null @@ -1,141 +0,0 @@ -/** - * @file llv4math.h - * @brief LLV4* class header file - vector processor enabled math - * - * $LicenseInfo:firstyear=2007&license=viewerlgpl$ - * Second Life Viewer Source Code - * Copyright (C) 2010, Linden Research, Inc. - * - * This library is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; - * version 2.1 of the License only. - * - * This library is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with this library; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - * - * Linden Research, Inc., 945 Battery Street, San Francisco, CA 94111 USA - * $/LicenseInfo$ - */ - -#ifndef LL_LLV4MATH_H -#define LL_LLV4MATH_H - -// *NOTE: We do not support SSE acceleration on Windows builds. -// Our minimum specification for the viewer includes 1 GHz Athlon processors, -// which covers the Athlon Thunderbird series that does not support SSE. -// -// Our header files include statements like this -// const F32 HAVOK_TIMESTEP = 1.f / 45.f; -// This creates "globals" that are included in each .obj file. If a single -// .cpp file has SSE code generation turned on (eg, llviewerjointmesh_sse.cpp) -// these globals will be initialized using SSE instructions. This causes SL -// to crash before main() on processors without SSE. Untangling all these -// headers/variables is too much work for the small performance gains of -// vectorization. -// -// Therefore we only support vectorization on builds where the everything is -// built with SSE or Altivec. See https://jira.secondlife.com/browse/VWR-1610 -// and https://jira.lindenlab.com/browse/SL-47720 for details. -// -// Sorry the code is such a mess. JC - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4MATH - GNUC -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -#if LL_GNUC && __GNUC__ >= 4 && __SSE__ - -#define LL_VECTORIZE 1 - -#if LL_DARWIN - -#include <Accelerate/Accelerate.h> -#include <xmmintrin.h> -typedef vFloat V4F32; - -#else - -#include <xmmintrin.h> -typedef float V4F32 __attribute__((vector_size(16))); - -#endif - -#endif -#if LL_GNUC - -#define LL_LLV4MATH_ALIGN_PREFIX -#define LL_LLV4MATH_ALIGN_POSTFIX __attribute__((aligned(16))) - -#endif - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4MATH - MSVC -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -// Only vectorize if the entire Windows build uses SSE. -// _M_IX86_FP is set when SSE code generation is turned on, and I have -// confirmed this in VS2003, VS2003 SP1, and VS2005. JC -#if LL_MSVC && _M_IX86_FP - -#define LL_VECTORIZE 1 - -#include <xmmintrin.h> - -typedef __m128 V4F32; - -#endif -#if LL_MSVC - -#define LL_LLV4MATH_ALIGN_PREFIX __declspec(align(16)) -#define LL_LLV4MATH_ALIGN_POSTFIX - -#endif - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4MATH - default - no vectorization -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -#if !LL_VECTORIZE - -#define LL_VECTORIZE 0 - -struct V4F32 { F32 __pad__[4]; }; - -inline F32 llv4lerp(F32 a, F32 b, F32 w) { return ( b - a ) * w + a; } - -#endif - -#ifndef LL_LLV4MATH_ALIGN_PREFIX -# define LL_LLV4MATH_ALIGN_PREFIX -#endif -#ifndef LL_LLV4MATH_ALIGN_POSTFIX -# define LL_LLV4MATH_ALIGN_POSTFIX -#endif - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4MATH -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - - -#define LLV4_NUM_AXIS 4 - -class LLV4Vector3; -class LLV4Matrix3; -class LLV4Matrix4; - -#endif diff --git a/indra/llmath/llv4matrix3.h b/indra/llmath/llv4matrix3.h deleted file mode 100644 index 270f5d7dae..0000000000 --- a/indra/llmath/llv4matrix3.h +++ /dev/null @@ -1,220 +0,0 @@ -/** - * @file llviewerjointmesh.cpp - * @brief LLV4* class header file - vector processor enabled math - * - * $LicenseInfo:firstyear=2007&license=viewerlgpl$ - * Second Life Viewer Source Code - * Copyright (C) 2010, Linden Research, Inc. - * - * This library is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; - * version 2.1 of the License only. - * - * This library is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with this library; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - * - * Linden Research, Inc., 945 Battery Street, San Francisco, CA 94111 USA - * $/LicenseInfo$ - */ - -#ifndef LL_LLV4MATRIX3_H -#define LL_LLV4MATRIX3_H - -#include "llv4math.h" -#include "llv4vector3.h" -#include "m3math.h" // for operator LLMatrix3() - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4Matrix3 -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -LL_LLV4MATH_ALIGN_PREFIX - -class LLV4Matrix3 -{ -public: - union { - F32 mMatrix[LLV4_NUM_AXIS][LLV4_NUM_AXIS]; - V4F32 mV[LLV4_NUM_AXIS]; - }; - - void lerp(const LLV4Matrix3 &a, const LLV4Matrix3 &b, const F32 &w); - void multiply(const LLVector3 &a, LLVector3& out) const; - void multiply(const LLVector4 &a, LLV4Vector3& out) const; - void multiply(const LLVector3 &a, LLV4Vector3& out) const; - - const LLV4Matrix3& transpose(); - const LLV4Matrix3& operator=(const LLMatrix3& a); - - operator LLMatrix3() const { return (reinterpret_cast<const LLMatrix4*>(const_cast<const F32*>(&mMatrix[0][0])))->getMat3(); } - - friend LLVector3 operator*(const LLVector3& a, const LLV4Matrix3& b); -} - -LL_LLV4MATH_ALIGN_POSTFIX; - - - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4Matrix3 - SSE -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -#if LL_VECTORIZE - -inline void LLV4Matrix3::lerp(const LLV4Matrix3 &a, const LLV4Matrix3 &b, const F32 &w) -{ - __m128 vw = _mm_set1_ps(w); - mV[VX] = _mm_add_ps(_mm_mul_ps(_mm_sub_ps(b.mV[VX], a.mV[VX]), vw), a.mV[VX]); // ( b - a ) * w + a - mV[VY] = _mm_add_ps(_mm_mul_ps(_mm_sub_ps(b.mV[VY], a.mV[VY]), vw), a.mV[VY]); - mV[VZ] = _mm_add_ps(_mm_mul_ps(_mm_sub_ps(b.mV[VZ], a.mV[VZ]), vw), a.mV[VZ]); -} - -inline void LLV4Matrix3::multiply(const LLVector3 &a, LLVector3& o) const -{ - LLV4Vector3 j; - j.v = _mm_mul_ps(_mm_set1_ps(a.mV[VX]), mV[VX]); // ( ax * vx ) + ... - j.v = _mm_add_ps(j.v , _mm_mul_ps(_mm_set1_ps(a.mV[VY]), mV[VY])); - j.v = _mm_add_ps(j.v , _mm_mul_ps(_mm_set1_ps(a.mV[VZ]), mV[VZ])); - o.setVec(j.mV); -} - -inline void LLV4Matrix3::multiply(const LLVector4 &a, LLV4Vector3& o) const -{ - o.v = _mm_mul_ps(_mm_set1_ps(a.mV[VX]), mV[VX]); // ( ax * vx ) + ... - o.v = _mm_add_ps(o.v , _mm_mul_ps(_mm_set1_ps(a.mV[VY]), mV[VY])); - o.v = _mm_add_ps(o.v , _mm_mul_ps(_mm_set1_ps(a.mV[VZ]), mV[VZ])); -} - -inline void LLV4Matrix3::multiply(const LLVector3 &a, LLV4Vector3& o) const -{ - o.v = _mm_mul_ps(_mm_set1_ps(a.mV[VX]), mV[VX]); // ( ax * vx ) + ... - o.v = _mm_add_ps(o.v , _mm_mul_ps(_mm_set1_ps(a.mV[VY]), mV[VY])); - o.v = _mm_add_ps(o.v , _mm_mul_ps(_mm_set1_ps(a.mV[VZ]), mV[VZ])); -} - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4Matrix3 -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -#else - -inline void LLV4Matrix3::lerp(const LLV4Matrix3 &a, const LLV4Matrix3 &b, const F32 &w) -{ - mMatrix[VX][VX] = llv4lerp(a.mMatrix[VX][VX], b.mMatrix[VX][VX], w); - mMatrix[VX][VY] = llv4lerp(a.mMatrix[VX][VY], b.mMatrix[VX][VY], w); - mMatrix[VX][VZ] = llv4lerp(a.mMatrix[VX][VZ], b.mMatrix[VX][VZ], w); - - mMatrix[VY][VX] = llv4lerp(a.mMatrix[VY][VX], b.mMatrix[VY][VX], w); - mMatrix[VY][VY] = llv4lerp(a.mMatrix[VY][VY], b.mMatrix[VY][VY], w); - mMatrix[VY][VZ] = llv4lerp(a.mMatrix[VY][VZ], b.mMatrix[VY][VZ], w); - - mMatrix[VZ][VX] = llv4lerp(a.mMatrix[VZ][VX], b.mMatrix[VZ][VX], w); - mMatrix[VZ][VY] = llv4lerp(a.mMatrix[VZ][VY], b.mMatrix[VZ][VY], w); - mMatrix[VZ][VZ] = llv4lerp(a.mMatrix[VZ][VZ], b.mMatrix[VZ][VZ], w); -} - -inline void LLV4Matrix3::multiply(const LLVector3 &a, LLVector3& o) const -{ - o.setVec( a.mV[VX] * mMatrix[VX][VX] + - a.mV[VY] * mMatrix[VY][VX] + - a.mV[VZ] * mMatrix[VZ][VX], - - a.mV[VX] * mMatrix[VX][VY] + - a.mV[VY] * mMatrix[VY][VY] + - a.mV[VZ] * mMatrix[VZ][VY], - - a.mV[VX] * mMatrix[VX][VZ] + - a.mV[VY] * mMatrix[VY][VZ] + - a.mV[VZ] * mMatrix[VZ][VZ]); -} - -inline void LLV4Matrix3::multiply(const LLVector4 &a, LLV4Vector3& o) const -{ - o.setVec( a.mV[VX] * mMatrix[VX][VX] + - a.mV[VY] * mMatrix[VY][VX] + - a.mV[VZ] * mMatrix[VZ][VX], - - a.mV[VX] * mMatrix[VX][VY] + - a.mV[VY] * mMatrix[VY][VY] + - a.mV[VZ] * mMatrix[VZ][VY], - - a.mV[VX] * mMatrix[VX][VZ] + - a.mV[VY] * mMatrix[VY][VZ] + - a.mV[VZ] * mMatrix[VZ][VZ]); -} - -inline void LLV4Matrix3::multiply(const LLVector3 &a, LLV4Vector3& o) const -{ - o.setVec( a.mV[VX] * mMatrix[VX][VX] + - a.mV[VY] * mMatrix[VY][VX] + - a.mV[VZ] * mMatrix[VZ][VX], - - a.mV[VX] * mMatrix[VX][VY] + - a.mV[VY] * mMatrix[VY][VY] + - a.mV[VZ] * mMatrix[VZ][VY], - - a.mV[VX] * mMatrix[VX][VZ] + - a.mV[VY] * mMatrix[VY][VZ] + - a.mV[VZ] * mMatrix[VZ][VZ]); -} - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4Matrix3 -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -#endif - -inline const LLV4Matrix3& LLV4Matrix3::transpose() -{ -#if LL_VECTORIZE && defined(_MM_TRANSPOSE4_PS) - _MM_TRANSPOSE4_PS(mV[VX], mV[VY], mV[VZ], mV[VW]); - return *this; -#else - F32 temp; - temp = mMatrix[VX][VY]; mMatrix[VX][VY] = mMatrix[VY][VX]; mMatrix[VY][VX] = temp; - temp = mMatrix[VX][VZ]; mMatrix[VX][VZ] = mMatrix[VZ][VX]; mMatrix[VZ][VX] = temp; - temp = mMatrix[VY][VZ]; mMatrix[VY][VZ] = mMatrix[VZ][VY]; mMatrix[VZ][VY] = temp; -#endif - return *this; -} - -inline const LLV4Matrix3& LLV4Matrix3::operator=(const LLMatrix3& a) -{ - memcpy(mMatrix[VX], a.mMatrix[VX], sizeof(F32) * 3 ); - memcpy(mMatrix[VY], a.mMatrix[VY], sizeof(F32) * 3 ); - memcpy(mMatrix[VZ], a.mMatrix[VZ], sizeof(F32) * 3 ); - return *this; -} - -inline LLVector3 operator*(const LLVector3& a, const LLV4Matrix3& b) -{ - return LLVector3( - a.mV[VX] * b.mMatrix[VX][VX] + - a.mV[VY] * b.mMatrix[VY][VX] + - a.mV[VZ] * b.mMatrix[VZ][VX], - - a.mV[VX] * b.mMatrix[VX][VY] + - a.mV[VY] * b.mMatrix[VY][VY] + - a.mV[VZ] * b.mMatrix[VZ][VY], - - a.mV[VX] * b.mMatrix[VX][VZ] + - a.mV[VY] * b.mMatrix[VY][VZ] + - a.mV[VZ] * b.mMatrix[VZ][VZ] ); -} - -#endif diff --git a/indra/llmath/llv4matrix4.h b/indra/llmath/llv4matrix4.h deleted file mode 100644 index 2eb49d9294..0000000000 --- a/indra/llmath/llv4matrix4.h +++ /dev/null @@ -1,249 +0,0 @@ -/** - * @file llviewerjointmesh.cpp - * @brief LLV4* class header file - vector processor enabled math - * - * $LicenseInfo:firstyear=2007&license=viewerlgpl$ - * Second Life Viewer Source Code - * Copyright (C) 2010, Linden Research, Inc. - * - * This library is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; - * version 2.1 of the License only. - * - * This library is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with this library; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - * - * Linden Research, Inc., 945 Battery Street, San Francisco, CA 94111 USA - * $/LicenseInfo$ - */ - -#ifndef LL_LLV4MATRIX4_H -#define LL_LLV4MATRIX4_H - -#include "llv4math.h" -#include "llv4matrix3.h" // just for operator LLV4Matrix3() -#include "llv4vector3.h" - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4Matrix4 -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -LL_LLV4MATH_ALIGN_PREFIX - -class LLV4Matrix4 -{ -public: - union { - F32 mMatrix[LLV4_NUM_AXIS][LLV4_NUM_AXIS]; - V4F32 mV[LLV4_NUM_AXIS]; - }; - - void lerp(const LLV4Matrix4 &a, const LLV4Matrix4 &b, const F32 &w); - void multiply(const LLVector3 &a, LLVector3& o) const; - void multiply(const LLVector3 &a, LLV4Vector3& o) const; - - const LLV4Matrix4& transpose(); - const LLV4Matrix4& translate(const LLVector3 &vec); - const LLV4Matrix4& translate(const LLV4Vector3 &vec); - const LLV4Matrix4& operator=(const LLMatrix4& a); - - operator LLMatrix4() const { return *(reinterpret_cast<const LLMatrix4*>(const_cast<const F32*>(&mMatrix[0][0]))); } - operator LLV4Matrix3() const { return *(reinterpret_cast<const LLV4Matrix3*>(const_cast<const F32*>(&mMatrix[0][0]))); } - - friend LLVector3 operator*(const LLVector3 &a, const LLV4Matrix4 &b); -} - -LL_LLV4MATH_ALIGN_POSTFIX; - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4Matrix4 - SSE -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -#if LL_VECTORIZE - -inline void LLV4Matrix4::lerp(const LLV4Matrix4 &a, const LLV4Matrix4 &b, const F32 &w) -{ - __m128 vw = _mm_set1_ps(w); - mV[VX] = _mm_add_ps(_mm_mul_ps(_mm_sub_ps(b.mV[VX], a.mV[VX]), vw), a.mV[VX]); // ( b - a ) * w + a - mV[VY] = _mm_add_ps(_mm_mul_ps(_mm_sub_ps(b.mV[VY], a.mV[VY]), vw), a.mV[VY]); - mV[VZ] = _mm_add_ps(_mm_mul_ps(_mm_sub_ps(b.mV[VZ], a.mV[VZ]), vw), a.mV[VZ]); - mV[VW] = _mm_add_ps(_mm_mul_ps(_mm_sub_ps(b.mV[VW], a.mV[VW]), vw), a.mV[VW]); -} - -inline void LLV4Matrix4::multiply(const LLVector3 &a, LLVector3& o) const -{ - LLV4Vector3 j; - j.v = _mm_add_ps(mV[VW], _mm_mul_ps(_mm_set1_ps(a.mV[VX]), mV[VX])); // ( ax * vx ) + vw - j.v = _mm_add_ps(j.v , _mm_mul_ps(_mm_set1_ps(a.mV[VY]), mV[VY])); - j.v = _mm_add_ps(j.v , _mm_mul_ps(_mm_set1_ps(a.mV[VZ]), mV[VZ])); - o.setVec(j.mV); -} - -inline void LLV4Matrix4::multiply(const LLVector3 &a, LLV4Vector3& o) const -{ - o.v = _mm_add_ps(mV[VW], _mm_mul_ps(_mm_set1_ps(a.mV[VX]), mV[VX])); // ( ax * vx ) + vw - o.v = _mm_add_ps(o.v , _mm_mul_ps(_mm_set1_ps(a.mV[VY]), mV[VY])); - o.v = _mm_add_ps(o.v , _mm_mul_ps(_mm_set1_ps(a.mV[VZ]), mV[VZ])); -} - -inline const LLV4Matrix4& LLV4Matrix4::translate(const LLV4Vector3 &vec) -{ - mV[VW] = _mm_add_ps(mV[VW], vec.v); - return (*this); -} - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4Matrix4 -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -#else - -inline void LLV4Matrix4::lerp(const LLV4Matrix4 &a, const LLV4Matrix4 &b, const F32 &w) -{ - mMatrix[VX][VX] = llv4lerp(a.mMatrix[VX][VX], b.mMatrix[VX][VX], w); - mMatrix[VX][VY] = llv4lerp(a.mMatrix[VX][VY], b.mMatrix[VX][VY], w); - mMatrix[VX][VZ] = llv4lerp(a.mMatrix[VX][VZ], b.mMatrix[VX][VZ], w); - - mMatrix[VY][VX] = llv4lerp(a.mMatrix[VY][VX], b.mMatrix[VY][VX], w); - mMatrix[VY][VY] = llv4lerp(a.mMatrix[VY][VY], b.mMatrix[VY][VY], w); - mMatrix[VY][VZ] = llv4lerp(a.mMatrix[VY][VZ], b.mMatrix[VY][VZ], w); - - mMatrix[VZ][VX] = llv4lerp(a.mMatrix[VZ][VX], b.mMatrix[VZ][VX], w); - mMatrix[VZ][VY] = llv4lerp(a.mMatrix[VZ][VY], b.mMatrix[VZ][VY], w); - mMatrix[VZ][VZ] = llv4lerp(a.mMatrix[VZ][VZ], b.mMatrix[VZ][VZ], w); - - mMatrix[VW][VX] = llv4lerp(a.mMatrix[VW][VX], b.mMatrix[VW][VX], w); - mMatrix[VW][VY] = llv4lerp(a.mMatrix[VW][VY], b.mMatrix[VW][VY], w); - mMatrix[VW][VZ] = llv4lerp(a.mMatrix[VW][VZ], b.mMatrix[VW][VZ], w); -} - -inline void LLV4Matrix4::multiply(const LLVector3 &a, LLVector3& o) const -{ - o.setVec( a.mV[VX] * mMatrix[VX][VX] + - a.mV[VY] * mMatrix[VY][VX] + - a.mV[VZ] * mMatrix[VZ][VX] + - mMatrix[VW][VX], - - a.mV[VX] * mMatrix[VX][VY] + - a.mV[VY] * mMatrix[VY][VY] + - a.mV[VZ] * mMatrix[VZ][VY] + - mMatrix[VW][VY], - - a.mV[VX] * mMatrix[VX][VZ] + - a.mV[VY] * mMatrix[VY][VZ] + - a.mV[VZ] * mMatrix[VZ][VZ] + - mMatrix[VW][VZ]); -} - -inline void LLV4Matrix4::multiply(const LLVector3 &a, LLV4Vector3& o) const -{ - o.setVec( a.mV[VX] * mMatrix[VX][VX] + - a.mV[VY] * mMatrix[VY][VX] + - a.mV[VZ] * mMatrix[VZ][VX] + - mMatrix[VW][VX], - - a.mV[VX] * mMatrix[VX][VY] + - a.mV[VY] * mMatrix[VY][VY] + - a.mV[VZ] * mMatrix[VZ][VY] + - mMatrix[VW][VY], - - a.mV[VX] * mMatrix[VX][VZ] + - a.mV[VY] * mMatrix[VY][VZ] + - a.mV[VZ] * mMatrix[VZ][VZ] + - mMatrix[VW][VZ]); -} - -inline const LLV4Matrix4& LLV4Matrix4::translate(const LLV4Vector3 &vec) -{ - mMatrix[3][0] += vec.mV[0]; - mMatrix[3][1] += vec.mV[1]; - mMatrix[3][2] += vec.mV[2]; - return (*this); -} - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4Matrix4 -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -#endif - -inline const LLV4Matrix4& LLV4Matrix4::operator=(const LLMatrix4& a) -{ - memcpy(mMatrix, a.mMatrix, sizeof(F32) * 16 ); - return *this; -} - -inline const LLV4Matrix4& LLV4Matrix4::transpose() -{ -#if LL_VECTORIZE && defined(_MM_TRANSPOSE4_PS) - _MM_TRANSPOSE4_PS(mV[VX], mV[VY], mV[VZ], mV[VW]); -#else - LLV4Matrix4 mat; - mat.mMatrix[0][0] = mMatrix[0][0]; - mat.mMatrix[1][0] = mMatrix[0][1]; - mat.mMatrix[2][0] = mMatrix[0][2]; - mat.mMatrix[3][0] = mMatrix[0][3]; - - mat.mMatrix[0][1] = mMatrix[1][0]; - mat.mMatrix[1][1] = mMatrix[1][1]; - mat.mMatrix[2][1] = mMatrix[1][2]; - mat.mMatrix[3][1] = mMatrix[1][3]; - - mat.mMatrix[0][2] = mMatrix[2][0]; - mat.mMatrix[1][2] = mMatrix[2][1]; - mat.mMatrix[2][2] = mMatrix[2][2]; - mat.mMatrix[3][2] = mMatrix[2][3]; - - mat.mMatrix[0][3] = mMatrix[3][0]; - mat.mMatrix[1][3] = mMatrix[3][1]; - mat.mMatrix[2][3] = mMatrix[3][2]; - mat.mMatrix[3][3] = mMatrix[3][3]; - - *this = mat; -#endif - return *this; -} - -inline const LLV4Matrix4& LLV4Matrix4::translate(const LLVector3 &vec) -{ - mMatrix[3][0] += vec.mV[0]; - mMatrix[3][1] += vec.mV[1]; - mMatrix[3][2] += vec.mV[2]; - return (*this); -} - -inline LLVector3 operator*(const LLVector3 &a, const LLV4Matrix4 &b) -{ - return LLVector3(a.mV[VX] * b.mMatrix[VX][VX] + - a.mV[VY] * b.mMatrix[VY][VX] + - a.mV[VZ] * b.mMatrix[VZ][VX] + - b.mMatrix[VW][VX], - - a.mV[VX] * b.mMatrix[VX][VY] + - a.mV[VY] * b.mMatrix[VY][VY] + - a.mV[VZ] * b.mMatrix[VZ][VY] + - b.mMatrix[VW][VY], - - a.mV[VX] * b.mMatrix[VX][VZ] + - a.mV[VY] * b.mMatrix[VY][VZ] + - a.mV[VZ] * b.mMatrix[VZ][VZ] + - b.mMatrix[VW][VZ]); -} - - -#endif diff --git a/indra/llmath/llv4vector3.h b/indra/llmath/llv4vector3.h deleted file mode 100644 index a340d53f5a..0000000000 --- a/indra/llmath/llv4vector3.h +++ /dev/null @@ -1,80 +0,0 @@ -/** - * @file llviewerjointmesh.cpp - * @brief LLV4* class header file - vector processor enabled math - * - * $LicenseInfo:firstyear=2007&license=viewerlgpl$ - * Second Life Viewer Source Code - * Copyright (C) 2010, Linden Research, Inc. - * - * This library is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; - * version 2.1 of the License only. - * - * This library is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with this library; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - * - * Linden Research, Inc., 945 Battery Street, San Francisco, CA 94111 USA - * $/LicenseInfo$ - */ - -#ifndef LL_LLV4VECTOR3_H -#define LL_LLV4VECTOR3_H - -#include "llv4math.h" - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4Vector3 -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -LL_LLV4MATH_ALIGN_PREFIX - -class LLV4Vector3 -{ -public: - union { - F32 mV[LLV4_NUM_AXIS]; - V4F32 v; - }; - - enum { - ALIGNMENT = 16 - }; - - void setVec(F32 x, F32 y, F32 z); - void setVec(F32 a); -} - -LL_LLV4MATH_ALIGN_POSTFIX; - -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- -// LLV4Vector3 -//----------------------------------------------------------------------------- -//----------------------------------------------------------------------------- - -inline void LLV4Vector3::setVec(F32 x, F32 y, F32 z) -{ - mV[VX] = x; - mV[VY] = y; - mV[VZ] = z; -} - -inline void LLV4Vector3::setVec(F32 a) -{ -#if LL_VECTORIZE - v = _mm_set1_ps(a); -#else - setVec(a, a, a); -#endif -} - -#endif |