/* Copyright (C) 2006, 2007 Sony Computer Entertainment Inc. All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. * Neither the name of the Sony Computer Entertainment Inc nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #ifndef _FLOATINVEC_H #define _FLOATINVEC_H #include #include #include #include #include "../c/vec_types.h" #undef bool namespace Vectormath { class boolInVec; //-------------------------------------------------------------------------------------------------- // floatInVec class // class floatInVec { private: vec_float4 mData; inline floatInVec(vec_float4 vec); public: inline floatInVec() {} // matches standard type conversions // inline floatInVec(boolInVec vec); // construct from a slot of vec_float4 // inline floatInVec(vec_float4 vec, int slot); // explicit cast from float // explicit inline floatInVec(float scalar); #ifdef _VECTORMATH_NO_SCALAR_CAST // explicit cast to float // inline float getAsFloat() const; #else // implicit cast to float // inline operator float() const; #endif // get vector data // float value is splatted across all word slots of vector // inline vec_float4 get128() const; // operators // inline const floatInVec operator ++ (int); inline const floatInVec operator -- (int); inline floatInVec& operator ++ (); inline floatInVec& operator -- (); inline const floatInVec operator - () const; inline floatInVec& operator = (floatInVec vec); inline floatInVec& operator *= (floatInVec vec); inline floatInVec& operator /= (floatInVec vec); inline floatInVec& operator += (floatInVec vec); inline floatInVec& operator -= (floatInVec vec); // friend functions // friend inline const floatInVec operator * (floatInVec vec0, floatInVec vec1); friend inline const floatInVec operator / (floatInVec vec0, floatInVec vec1); friend inline const floatInVec operator + (floatInVec vec0, floatInVec vec1); friend inline const floatInVec operator - (floatInVec vec0, floatInVec vec1); friend inline const floatInVec select(floatInVec vec0, floatInVec vec1, boolInVec select_vec1); }; //-------------------------------------------------------------------------------------------------- // floatInVec functions // // operators // inline const floatInVec operator * (floatInVec vec0, floatInVec vec1); inline const floatInVec operator / (floatInVec vec0, floatInVec vec1); inline const floatInVec operator + (floatInVec vec0, floatInVec vec1); inline const floatInVec operator - (floatInVec vec0, floatInVec vec1); inline const boolInVec operator < (floatInVec vec0, floatInVec vec1); inline const boolInVec operator <= (floatInVec vec0, floatInVec vec1); inline const boolInVec operator > (floatInVec vec0, floatInVec vec1); inline const boolInVec operator >= (floatInVec vec0, floatInVec vec1); inline const boolInVec operator == (floatInVec vec0, floatInVec vec1); inline const boolInVec operator != (floatInVec vec0, floatInVec vec1); // select between vec0 and vec1 using boolInVec. // false selects vec0, true selects vec1 // inline const floatInVec select(floatInVec vec0, floatInVec vec1, boolInVec select_vec1); } // namespace Vectormath //-------------------------------------------------------------------------------------------------- // floatInVec implementation // #include "boolInVec.h" namespace Vectormath { inline floatInVec::floatInVec(vec_float4 vec) { mData = vec; } inline floatInVec::floatInVec(boolInVec vec) { mData = vec_ctf(vec_sub((vec_uint4){0,0,0,0}, vec.get128()), 0); } inline floatInVec::floatInVec(vec_float4 vec, int slot) { #ifdef __GNUC__ if (__builtin_constant_p(slot)) { mData = vec_splat(vec, slot); } else #endif { const vec_uchar16 shiftpattern = vec_lvsl(0, (float *)(size_t)(slot << 2)); mData = vec_splat(vec_perm(vec, vec, shiftpattern), 0); } } inline floatInVec::floatInVec(float scalar) { #ifdef __GNUC__ if (__builtin_constant_p(scalar)) { mData = (vec_float4){scalar, scalar, scalar, scalar}; } else #endif { vec_float4 vec = vec_ld(0, &scalar); mData = vec_splat(vec_perm(vec, vec, vec_lvsl(0, &scalar)), 0); } } #ifdef _VECTORMATH_NO_SCALAR_CAST inline float floatInVec::getAsFloat() const #else inline floatInVec::operator float() const #endif { return *((float *)&mData); } inline vec_float4 floatInVec::get128() const { return mData; } inline const floatInVec floatInVec::operator ++ (int) { vec_float4 olddata = mData; operator ++(); return floatInVec(olddata); } inline const floatInVec floatInVec::operator -- (int) { vec_float4 olddata = mData; operator --(); return floatInVec(olddata); } inline floatInVec& floatInVec::operator ++ () { *this += floatInVec((vec_float4){1.0f,1.0f,1.0f,1.0f}); return *this; } inline floatInVec& floatInVec::operator -- () { *this -= floatInVec((vec_float4){1.0f,1.0f,1.0f,1.0f}); return *this; } inline const floatInVec floatInVec::operator - () const { return floatInVec((vec_float4)vec_xor((vec_uint4)mData, (vec_uint4){0x80000000,0x80000000,0x80000000,0x80000000})); } inline floatInVec& floatInVec::operator = (floatInVec vec) { mData = vec.mData; return *this; } inline floatInVec& floatInVec::operator *= (floatInVec vec) { *this = *this * vec; return *this; } inline floatInVec& floatInVec::operator /= (floatInVec vec) { *this = *this / vec; return *this; } inline floatInVec& floatInVec::operator += (floatInVec vec) { *this = *this + vec; return *this; } inline floatInVec& floatInVec::operator -= (floatInVec vec) { *this = *this - vec; return *this; } inline const floatInVec operator * (floatInVec vec0, floatInVec vec1) { return floatInVec(vec_madd(vec0.get128(), vec1.get128(), (vec_float4){0,0,0,0})); } inline const floatInVec operator / (floatInVec num, floatInVec den) { return floatInVec(divf4(num.get128(), den.get128())); } inline const floatInVec operator + (floatInVec vec0, floatInVec vec1) { return floatInVec(vec_add(vec0.get128(), vec1.get128())); } inline const floatInVec operator - (floatInVec vec0, floatInVec vec1) { return floatInVec(vec_sub(vec0.get128(), vec1.get128())); } inline const boolInVec operator < (floatInVec vec0, floatInVec vec1) { return boolInVec((vec_uint4)vec_cmpgt(vec1.get128(), vec0.get128())); } inline const boolInVec operator <= (floatInVec vec0, floatInVec vec1) { return !(vec0 > vec1); } inline const boolInVec operator > (floatInVec vec0, floatInVec vec1) { return boolInVec((vec_uint4)vec_cmpgt(vec0.get128(), vec1.get128())); } inline const boolInVec operator >= (floatInVec vec0, floatInVec vec1) { return !(vec0 < vec1); } inline const boolInVec operator == (floatInVec vec0, floatInVec vec1) { return boolInVec((vec_uint4)vec_cmpeq(vec0.get128(), vec1.get128())); } inline const boolInVec operator != (floatInVec vec0, floatInVec vec1) { return !(vec0 == vec1); } inline const floatInVec select(floatInVec vec0, floatInVec vec1, boolInVec select_vec1) { return floatInVec(vec_sel(vec0.get128(), vec1.get128(), select_vec1.get128())); } } // namespace Vectormath #endif // floatInVec_h