diff options
Diffstat (limited to 'extern/bullet2/src/LinearMath/btVector3.h')
| -rw-r--r-- | extern/bullet2/src/LinearMath/btVector3.h | 808 |
1 files changed, 692 insertions, 116 deletions
diff --git a/extern/bullet2/src/LinearMath/btVector3.h b/extern/bullet2/src/LinearMath/btVector3.h index d99b7c83ae3..1cf65358803 100644 --- a/extern/bullet2/src/LinearMath/btVector3.h +++ b/extern/bullet2/src/LinearMath/btVector3.h @@ -17,9 +17,10 @@ subject to the following restrictions: #ifndef BT_VECTOR3_H #define BT_VECTOR3_H - +//#include <stdint.h> #include "btScalar.h" #include "btMinMax.h" +#include "btAlignedAllocator.h" #ifdef BT_USE_DOUBLE_PRECISION #define btVector3Data btVector3DoubleData @@ -29,8 +30,46 @@ subject to the following restrictions: #define btVector3DataName "btVector3FloatData" #endif //BT_USE_DOUBLE_PRECISION +#if defined BT_USE_SSE + +//typedef uint32_t __m128i __attribute__ ((vector_size(16))); + +#ifdef _MSC_VER +#pragma warning(disable: 4556) // value of intrinsic immediate argument '4294967239' is out of range '0 - 255' +#endif + + +#define BT_SHUFFLE(x,y,z,w) ((w)<<6 | (z)<<4 | (y)<<2 | (x)) +//#define bt_pshufd_ps( _a, _mask ) (__m128) _mm_shuffle_epi32((__m128i)(_a), (_mask) ) +#define bt_pshufd_ps( _a, _mask ) _mm_shuffle_ps((_a), (_a), (_mask) ) +#define bt_splat3_ps( _a, _i ) bt_pshufd_ps((_a), BT_SHUFFLE(_i,_i,_i, 3) ) +#define bt_splat_ps( _a, _i ) bt_pshufd_ps((_a), BT_SHUFFLE(_i,_i,_i,_i) ) + +#define btv3AbsiMask (_mm_set_epi32(0x00000000, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF)) +#define btvAbsMask (_mm_set_epi32( 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF)) +#define btvFFF0Mask (_mm_set_epi32(0x00000000, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF)) +#define btv3AbsfMask btCastiTo128f(btv3AbsiMask) +#define btvFFF0fMask btCastiTo128f(btvFFF0Mask) +#define btvxyzMaskf btvFFF0fMask +#define btvAbsfMask btCastiTo128f(btvAbsMask) + + + +const __m128 ATTRIBUTE_ALIGNED16(btvMzeroMask) = {-0.0f, -0.0f, -0.0f, -0.0f}; +const __m128 ATTRIBUTE_ALIGNED16(v1110) = {1.0f, 1.0f, 1.0f, 0.0f}; +const __m128 ATTRIBUTE_ALIGNED16(vHalf) = {0.5f, 0.5f, 0.5f, 0.5f}; +const __m128 ATTRIBUTE_ALIGNED16(v1_5) = {1.5f, 1.5f, 1.5f, 1.5f}; + +#endif +#ifdef BT_USE_NEON +const float32x4_t ATTRIBUTE_ALIGNED16(btvMzeroMask) = (float32x4_t){-0.0f, -0.0f, -0.0f, -0.0f}; +const int32x4_t ATTRIBUTE_ALIGNED16(btvFFF0Mask) = (int32x4_t){0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x0}; +const int32x4_t ATTRIBUTE_ALIGNED16(btvAbsMask) = (int32x4_t){0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF}; +const int32x4_t ATTRIBUTE_ALIGNED16(btv3AbsMask) = (int32x4_t){0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x0}; + +#endif /**@brief btVector3 can be used to represent 3D points and vectors. * It has an un-used w component to suit 16-byte alignment when btVector3 is stored in containers. This extra component can be used by derived classes (Quaternion?) or by user @@ -40,6 +79,8 @@ ATTRIBUTE_ALIGNED16(class) btVector3 { public: + BT_DECLARE_ALIGNED_ALLOCATOR(); + #if defined (__SPU__) && defined (__CELLOS_LV2__) btScalar m_floats[4]; public: @@ -49,28 +90,31 @@ public: } public: #else //__CELLOS_LV2__ __SPU__ -#ifdef BT_USE_SSE // _WIN32 - union { - __m128 mVec128; - btScalar m_floats[4]; - }; - SIMD_FORCE_INLINE __m128 get128() const - { - return mVec128; - } - SIMD_FORCE_INLINE void set128(__m128 v128) - { - mVec128 = v128; - } -#else - btScalar m_floats[4]; -#endif + #if defined (BT_USE_SSE) || defined(BT_USE_NEON) // _WIN32 || ARM + union { + btSimdFloat4 mVec128; + btScalar m_floats[4]; + }; + SIMD_FORCE_INLINE btSimdFloat4 get128() const + { + return mVec128; + } + SIMD_FORCE_INLINE void set128(btSimdFloat4 v128) + { + mVec128 = v128; + } + #else + btScalar m_floats[4]; + #endif #endif //__CELLOS_LV2__ __SPU__ public: /**@brief No initialization constructor */ - SIMD_FORCE_INLINE btVector3() {} + SIMD_FORCE_INLINE btVector3() + { + + } @@ -79,21 +123,50 @@ public: * @param y Y value * @param z Z value */ - SIMD_FORCE_INLINE btVector3(const btScalar& x, const btScalar& y, const btScalar& z) + SIMD_FORCE_INLINE btVector3(const btScalar& _x, const btScalar& _y, const btScalar& _z) { - m_floats[0] = x; - m_floats[1] = y; - m_floats[2] = z; - m_floats[3] = btScalar(0.); + m_floats[0] = _x; + m_floats[1] = _y; + m_floats[2] = _z; + m_floats[3] = btScalar(0.f); } - +#if (defined (BT_USE_SSE_IN_API) && defined (BT_USE_SSE) )|| defined (BT_USE_NEON) + // Set Vector + SIMD_FORCE_INLINE btVector3( btSimdFloat4 v) + { + mVec128 = v; + } + + // Copy constructor + SIMD_FORCE_INLINE btVector3(const btVector3& rhs) + { + mVec128 = rhs.mVec128; + } + + // Assignment Operator + SIMD_FORCE_INLINE btVector3& + operator=(const btVector3& v) + { + mVec128 = v.mVec128; + + return *this; + } +#endif // #if defined (BT_USE_SSE_IN_API) || defined (BT_USE_NEON) + /**@brief Add a vector to this one * @param The vector to add to this one */ SIMD_FORCE_INLINE btVector3& operator+=(const btVector3& v) { - - m_floats[0] += v.m_floats[0]; m_floats[1] += v.m_floats[1];m_floats[2] += v.m_floats[2]; +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + mVec128 = _mm_add_ps(mVec128, v.mVec128); +#elif defined(BT_USE_NEON) + mVec128 = vaddq_f32(mVec128, v.mVec128); +#else + m_floats[0] += v.m_floats[0]; + m_floats[1] += v.m_floats[1]; + m_floats[2] += v.m_floats[2]; +#endif return *this; } @@ -102,14 +175,33 @@ public: * @param The vector to subtract */ SIMD_FORCE_INLINE btVector3& operator-=(const btVector3& v) { - m_floats[0] -= v.m_floats[0]; m_floats[1] -= v.m_floats[1];m_floats[2] -= v.m_floats[2]; +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + mVec128 = _mm_sub_ps(mVec128, v.mVec128); +#elif defined(BT_USE_NEON) + mVec128 = vsubq_f32(mVec128, v.mVec128); +#else + m_floats[0] -= v.m_floats[0]; + m_floats[1] -= v.m_floats[1]; + m_floats[2] -= v.m_floats[2]; +#endif return *this; } + /**@brief Scale the vector * @param s Scale factor */ SIMD_FORCE_INLINE btVector3& operator*=(const btScalar& s) { - m_floats[0] *= s; m_floats[1] *= s;m_floats[2] *= s; +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + __m128 vs = _mm_load_ss(&s); // (S 0 0 0) + vs = bt_pshufd_ps(vs, 0x80); // (S S S 0.0) + mVec128 = _mm_mul_ps(mVec128, vs); +#elif defined(BT_USE_NEON) + mVec128 = vmulq_n_f32(mVec128, s); +#else + m_floats[0] *= s; + m_floats[1] *= s; + m_floats[2] *= s; +#endif return *this; } @@ -118,14 +210,42 @@ public: SIMD_FORCE_INLINE btVector3& operator/=(const btScalar& s) { btFullAssert(s != btScalar(0.0)); + +#if 0 //defined(BT_USE_SSE_IN_API) +// this code is not faster ! + __m128 vs = _mm_load_ss(&s); + vs = _mm_div_ss(v1110, vs); + vs = bt_pshufd_ps(vs, 0x00); // (S S S S) + + mVec128 = _mm_mul_ps(mVec128, vs); + + return *this; +#else return *this *= btScalar(1.0) / s; +#endif } /**@brief Return the dot product * @param v The other vector in the dot product */ SIMD_FORCE_INLINE btScalar dot(const btVector3& v) const { - return m_floats[0] * v.m_floats[0] + m_floats[1] * v.m_floats[1] +m_floats[2] * v.m_floats[2]; +#if defined BT_USE_SIMD_VECTOR3 && defined (BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + __m128 vd = _mm_mul_ps(mVec128, v.mVec128); + __m128 z = _mm_movehl_ps(vd, vd); + __m128 y = _mm_shuffle_ps(vd, vd, 0x55); + vd = _mm_add_ss(vd, y); + vd = _mm_add_ss(vd, z); + return _mm_cvtss_f32(vd); +#elif defined(BT_USE_NEON) + float32x4_t vd = vmulq_f32(mVec128, v.mVec128); + float32x2_t x = vpadd_f32(vget_low_f32(vd), vget_low_f32(vd)); + x = vadd_f32(x, vget_high_f32(vd)); + return vget_lane_f32(x, 0); +#else + return m_floats[0] * v.m_floats[0] + + m_floats[1] * v.m_floats[1] + + m_floats[2] * v.m_floats[2]; +#endif } /**@brief Return the length of the vector squared */ @@ -165,7 +285,44 @@ public: * x^2 + y^2 + z^2 = 1 */ SIMD_FORCE_INLINE btVector3& normalize() { +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + // dot product first + __m128 vd = _mm_mul_ps(mVec128, mVec128); + __m128 z = _mm_movehl_ps(vd, vd); + __m128 y = _mm_shuffle_ps(vd, vd, 0x55); + vd = _mm_add_ss(vd, y); + vd = _mm_add_ss(vd, z); + + #if 0 + vd = _mm_sqrt_ss(vd); + vd = _mm_div_ss(v1110, vd); + vd = bt_splat_ps(vd, 0x80); + mVec128 = _mm_mul_ps(mVec128, vd); + #else + + // NR step 1/sqrt(x) - vd is x, y is output + y = _mm_rsqrt_ss(vd); // estimate + + // one step NR + z = v1_5; + vd = _mm_mul_ss(vd, vHalf); // vd * 0.5 + //x2 = vd; + vd = _mm_mul_ss(vd, y); // vd * 0.5 * y0 + vd = _mm_mul_ss(vd, y); // vd * 0.5 * y0 * y0 + z = _mm_sub_ss(z, vd); // 1.5 - vd * 0.5 * y0 * y0 + + y = _mm_mul_ss(y, z); // y0 * (1.5 - vd * 0.5 * y0 * y0) + + y = bt_splat_ps(y, 0x80); + mVec128 = _mm_mul_ps(mVec128, y); + + #endif + + + return *this; +#else return *this /= length(); +#endif } /**@brief Return a normalized version of this vector */ @@ -184,29 +341,112 @@ public: btFullAssert(s != btScalar(0.0)); return btAcos(dot(v) / s); } + /**@brief Return a vector will the absolute values of each element */ SIMD_FORCE_INLINE btVector3 absolute() const { + +#if defined BT_USE_SIMD_VECTOR3 && defined (BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + return btVector3(_mm_and_ps(mVec128, btv3AbsfMask)); +#elif defined(BT_USE_NEON) + return btVector3(vabsq_f32(mVec128)); +#else return btVector3( btFabs(m_floats[0]), btFabs(m_floats[1]), btFabs(m_floats[2])); +#endif } + /**@brief Return the cross product between this and another vector * @param v The other vector */ SIMD_FORCE_INLINE btVector3 cross(const btVector3& v) const { +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + __m128 T, V; + + T = bt_pshufd_ps(mVec128, BT_SHUFFLE(1, 2, 0, 3)); // (Y Z X 0) + V = bt_pshufd_ps(v.mVec128, BT_SHUFFLE(1, 2, 0, 3)); // (Y Z X 0) + + V = _mm_mul_ps(V, mVec128); + T = _mm_mul_ps(T, v.mVec128); + V = _mm_sub_ps(V, T); + + V = bt_pshufd_ps(V, BT_SHUFFLE(1, 2, 0, 3)); + return btVector3(V); +#elif defined(BT_USE_NEON) + float32x4_t T, V; + // form (Y, Z, X, _) of mVec128 and v.mVec128 + float32x2_t Tlow = vget_low_f32(mVec128); + float32x2_t Vlow = vget_low_f32(v.mVec128); + T = vcombine_f32(vext_f32(Tlow, vget_high_f32(mVec128), 1), Tlow); + V = vcombine_f32(vext_f32(Vlow, vget_high_f32(v.mVec128), 1), Vlow); + + V = vmulq_f32(V, mVec128); + T = vmulq_f32(T, v.mVec128); + V = vsubq_f32(V, T); + Vlow = vget_low_f32(V); + // form (Y, Z, X, _); + V = vcombine_f32(vext_f32(Vlow, vget_high_f32(V), 1), Vlow); + V = (float32x4_t)vandq_s32((int32x4_t)V, btvFFF0Mask); + + return btVector3(V); +#else return btVector3( - m_floats[1] * v.m_floats[2] -m_floats[2] * v.m_floats[1], + m_floats[1] * v.m_floats[2] - m_floats[2] * v.m_floats[1], m_floats[2] * v.m_floats[0] - m_floats[0] * v.m_floats[2], m_floats[0] * v.m_floats[1] - m_floats[1] * v.m_floats[0]); +#endif } SIMD_FORCE_INLINE btScalar triple(const btVector3& v1, const btVector3& v2) const { - return m_floats[0] * (v1.m_floats[1] * v2.m_floats[2] - v1.m_floats[2] * v2.m_floats[1]) + +#if defined BT_USE_SIMD_VECTOR3 && defined (BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + // cross: + __m128 T = _mm_shuffle_ps(v1.mVec128, v1.mVec128, BT_SHUFFLE(1, 2, 0, 3)); // (Y Z X 0) + __m128 V = _mm_shuffle_ps(v2.mVec128, v2.mVec128, BT_SHUFFLE(1, 2, 0, 3)); // (Y Z X 0) + + V = _mm_mul_ps(V, v1.mVec128); + T = _mm_mul_ps(T, v2.mVec128); + V = _mm_sub_ps(V, T); + + V = _mm_shuffle_ps(V, V, BT_SHUFFLE(1, 2, 0, 3)); + + // dot: + V = _mm_mul_ps(V, mVec128); + __m128 z = _mm_movehl_ps(V, V); + __m128 y = _mm_shuffle_ps(V, V, 0x55); + V = _mm_add_ss(V, y); + V = _mm_add_ss(V, z); + return _mm_cvtss_f32(V); + +#elif defined(BT_USE_NEON) + // cross: + float32x4_t T, V; + // form (Y, Z, X, _) of mVec128 and v.mVec128 + float32x2_t Tlow = vget_low_f32(v1.mVec128); + float32x2_t Vlow = vget_low_f32(v2.mVec128); + T = vcombine_f32(vext_f32(Tlow, vget_high_f32(v1.mVec128), 1), Tlow); + V = vcombine_f32(vext_f32(Vlow, vget_high_f32(v2.mVec128), 1), Vlow); + + V = vmulq_f32(V, v1.mVec128); + T = vmulq_f32(T, v2.mVec128); + V = vsubq_f32(V, T); + Vlow = vget_low_f32(V); + // form (Y, Z, X, _); + V = vcombine_f32(vext_f32(Vlow, vget_high_f32(V), 1), Vlow); + + // dot: + V = vmulq_f32(mVec128, V); + float32x2_t x = vpadd_f32(vget_low_f32(V), vget_low_f32(V)); + x = vadd_f32(x, vget_high_f32(V)); + return vget_lane_f32(x, 0); +#else + return + m_floats[0] * (v1.m_floats[1] * v2.m_floats[2] - v1.m_floats[2] * v2.m_floats[1]) + m_floats[1] * (v1.m_floats[2] * v2.m_floats[0] - v1.m_floats[0] * v2.m_floats[2]) + m_floats[2] * (v1.m_floats[0] * v2.m_floats[1] - v1.m_floats[1] * v2.m_floats[0]); +#endif } /**@brief Return the axis with the smallest value @@ -233,14 +473,31 @@ public: return absolute().maxAxis(); } + SIMD_FORCE_INLINE void setInterpolate3(const btVector3& v0, const btVector3& v1, btScalar rt) { +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + __m128 vrt = _mm_load_ss(&rt); // (rt 0 0 0) + btScalar s = btScalar(1.0) - rt; + __m128 vs = _mm_load_ss(&s); // (S 0 0 0) + vs = bt_pshufd_ps(vs, 0x80); // (S S S 0.0) + __m128 r0 = _mm_mul_ps(v0.mVec128, vs); + vrt = bt_pshufd_ps(vrt, 0x80); // (rt rt rt 0.0) + __m128 r1 = _mm_mul_ps(v1.mVec128, vrt); + __m128 tmp3 = _mm_add_ps(r0,r1); + mVec128 = tmp3; +#elif defined(BT_USE_NEON) + mVec128 = vsubq_f32(v1.mVec128, v0.mVec128); + mVec128 = vmulq_n_f32(mVec128, rt); + mVec128 = vaddq_f32(mVec128, v0.mVec128); +#else btScalar s = btScalar(1.0) - rt; m_floats[0] = s * v0.m_floats[0] + rt * v1.m_floats[0]; m_floats[1] = s * v0.m_floats[1] + rt * v1.m_floats[1]; m_floats[2] = s * v0.m_floats[2] + rt * v1.m_floats[2]; //don't do the unused w component // m_co[3] = s * v0[3] + rt * v1[3]; +#endif } /**@brief Return the linear interpolation between this and another vector @@ -248,16 +505,41 @@ public: * @param t The ration of this to v (t = 0 => return this, t=1 => return other) */ SIMD_FORCE_INLINE btVector3 lerp(const btVector3& v, const btScalar& t) const { - return btVector3(m_floats[0] + (v.m_floats[0] - m_floats[0]) * t, - m_floats[1] + (v.m_floats[1] - m_floats[1]) * t, - m_floats[2] + (v.m_floats[2] -m_floats[2]) * t); +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + __m128 vt = _mm_load_ss(&t); // (t 0 0 0) + vt = bt_pshufd_ps(vt, 0x80); // (rt rt rt 0.0) + __m128 vl = _mm_sub_ps(v.mVec128, mVec128); + vl = _mm_mul_ps(vl, vt); + vl = _mm_add_ps(vl, mVec128); + + return btVector3(vl); +#elif defined(BT_USE_NEON) + float32x4_t vl = vsubq_f32(v.mVec128, mVec128); + vl = vmulq_n_f32(vl, t); + vl = vaddq_f32(vl, mVec128); + + return btVector3(vl); +#else + return + btVector3( m_floats[0] + (v.m_floats[0] - m_floats[0]) * t, + m_floats[1] + (v.m_floats[1] - m_floats[1]) * t, + m_floats[2] + (v.m_floats[2] - m_floats[2]) * t); +#endif } /**@brief Elementwise multiply this vector by the other * @param v The other vector */ SIMD_FORCE_INLINE btVector3& operator*=(const btVector3& v) { - m_floats[0] *= v.m_floats[0]; m_floats[1] *= v.m_floats[1];m_floats[2] *= v.m_floats[2]; +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + mVec128 = _mm_mul_ps(mVec128, v.mVec128); +#elif defined(BT_USE_NEON) + mVec128 = vmulq_f32(mVec128, v.mVec128); +#else + m_floats[0] *= v.m_floats[0]; + m_floats[1] *= v.m_floats[1]; + m_floats[2] *= v.m_floats[2]; +#endif return *this; } @@ -268,13 +550,13 @@ public: /**@brief Return the z value */ SIMD_FORCE_INLINE const btScalar& getZ() const { return m_floats[2]; } /**@brief Set the x value */ - SIMD_FORCE_INLINE void setX(btScalar x) { m_floats[0] = x;}; + SIMD_FORCE_INLINE void setX(btScalar _x) { m_floats[0] = _x;}; /**@brief Set the y value */ - SIMD_FORCE_INLINE void setY(btScalar y) { m_floats[1] = y;}; + SIMD_FORCE_INLINE void setY(btScalar _y) { m_floats[1] = _y;}; /**@brief Set the z value */ - SIMD_FORCE_INLINE void setZ(btScalar z) {m_floats[2] = z;}; + SIMD_FORCE_INLINE void setZ(btScalar _z) { m_floats[2] = _z;}; /**@brief Set the w value */ - SIMD_FORCE_INLINE void setW(btScalar w) { m_floats[3] = w;}; + SIMD_FORCE_INLINE void setW(btScalar _w) { m_floats[3] = _w;}; /**@brief Return the x value */ SIMD_FORCE_INLINE const btScalar& x() const { return m_floats[0]; } /**@brief Return the y value */ @@ -292,7 +574,14 @@ public: SIMD_FORCE_INLINE bool operator==(const btVector3& other) const { - return ((m_floats[3]==other.m_floats[3]) && (m_floats[2]==other.m_floats[2]) && (m_floats[1]==other.m_floats[1]) && (m_floats[0]==other.m_floats[0])); +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + return (0xf == _mm_movemask_ps((__m128)_mm_cmpeq_ps(mVec128, other.mVec128))); +#else + return ((m_floats[3]==other.m_floats[3]) && + (m_floats[2]==other.m_floats[2]) && + (m_floats[1]==other.m_floats[1]) && + (m_floats[0]==other.m_floats[0])); +#endif } SIMD_FORCE_INLINE bool operator!=(const btVector3& other) const @@ -300,103 +589,230 @@ public: return !(*this == other); } - /**@brief Set each element to the max of the current values and the values of another btVector3 + /**@brief Set each element to the max of the current values and the values of another btVector3 * @param other The other btVector3 to compare with */ - SIMD_FORCE_INLINE void setMax(const btVector3& other) - { - btSetMax(m_floats[0], other.m_floats[0]); - btSetMax(m_floats[1], other.m_floats[1]); - btSetMax(m_floats[2], other.m_floats[2]); - btSetMax(m_floats[3], other.w()); - } + SIMD_FORCE_INLINE void setMax(const btVector3& other) + { +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + mVec128 = _mm_max_ps(mVec128, other.mVec128); +#elif defined(BT_USE_NEON) + mVec128 = vmaxq_f32(mVec128, other.mVec128); +#else + btSetMax(m_floats[0], other.m_floats[0]); + btSetMax(m_floats[1], other.m_floats[1]); + btSetMax(m_floats[2], other.m_floats[2]); + btSetMax(m_floats[3], other.w()); +#endif + } + /**@brief Set each element to the min of the current values and the values of another btVector3 * @param other The other btVector3 to compare with */ - SIMD_FORCE_INLINE void setMin(const btVector3& other) - { - btSetMin(m_floats[0], other.m_floats[0]); - btSetMin(m_floats[1], other.m_floats[1]); - btSetMin(m_floats[2], other.m_floats[2]); - btSetMin(m_floats[3], other.w()); - } - - SIMD_FORCE_INLINE void setValue(const btScalar& x, const btScalar& y, const btScalar& z) - { - m_floats[0]=x; - m_floats[1]=y; - m_floats[2]=z; - m_floats[3] = btScalar(0.); - } - - void getSkewSymmetricMatrix(btVector3* v0,btVector3* v1,btVector3* v2) const - { - v0->setValue(0. ,-z() ,y()); - v1->setValue(z() ,0. ,-x()); - v2->setValue(-y() ,x() ,0.); - } - - void setZero() - { - setValue(btScalar(0.),btScalar(0.),btScalar(0.)); - } - - SIMD_FORCE_INLINE bool isZero() const - { - return m_floats[0] == btScalar(0) && m_floats[1] == btScalar(0) && m_floats[2] == btScalar(0); - } - - SIMD_FORCE_INLINE bool fuzzyZero() const - { - return length2() < SIMD_EPSILON; - } - - SIMD_FORCE_INLINE void serialize(struct btVector3Data& dataOut) const; + SIMD_FORCE_INLINE void setMin(const btVector3& other) + { +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + mVec128 = _mm_min_ps(mVec128, other.mVec128); +#elif defined(BT_USE_NEON) + mVec128 = vminq_f32(mVec128, other.mVec128); +#else + btSetMin(m_floats[0], other.m_floats[0]); + btSetMin(m_floats[1], other.m_floats[1]); + btSetMin(m_floats[2], other.m_floats[2]); + btSetMin(m_floats[3], other.w()); +#endif + } - SIMD_FORCE_INLINE void deSerialize(const struct btVector3Data& dataIn); + SIMD_FORCE_INLINE void setValue(const btScalar& _x, const btScalar& _y, const btScalar& _z) + { + m_floats[0]=_x; + m_floats[1]=_y; + m_floats[2]=_z; + m_floats[3] = btScalar(0.f); + } - SIMD_FORCE_INLINE void serializeFloat(struct btVector3FloatData& dataOut) const; + void getSkewSymmetricMatrix(btVector3* v0,btVector3* v1,btVector3* v2) const + { +#if defined BT_USE_SIMD_VECTOR3 && defined (BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + + __m128 V = _mm_and_ps(mVec128, btvFFF0fMask); + __m128 V0 = _mm_xor_ps(btvMzeroMask, V); + __m128 V2 = _mm_movelh_ps(V0, V); + + __m128 V1 = _mm_shuffle_ps(V, V0, 0xCE); + + V0 = _mm_shuffle_ps(V0, V, 0xDB); + V2 = _mm_shuffle_ps(V2, V, 0xF9); + + v0->mVec128 = V0; + v1->mVec128 = V1; + v2->mVec128 = V2; +#else + v0->setValue(0. ,-z() ,y()); + v1->setValue(z() ,0. ,-x()); + v2->setValue(-y() ,x() ,0.); +#endif + } - SIMD_FORCE_INLINE void deSerializeFloat(const struct btVector3FloatData& dataIn); + void setZero() + { +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + mVec128 = (__m128)_mm_xor_ps(mVec128, mVec128); +#elif defined(BT_USE_NEON) + int32x4_t vi = vdupq_n_s32(0); + mVec128 = vreinterpretq_f32_s32(vi); +#else + setValue(btScalar(0.),btScalar(0.),btScalar(0.)); +#endif + } - SIMD_FORCE_INLINE void serializeDouble(struct btVector3DoubleData& dataOut) const; + SIMD_FORCE_INLINE bool isZero() const + { + return m_floats[0] == btScalar(0) && m_floats[1] == btScalar(0) && m_floats[2] == btScalar(0); + } - SIMD_FORCE_INLINE void deSerializeDouble(const struct btVector3DoubleData& dataIn); + SIMD_FORCE_INLINE bool fuzzyZero() const + { + return length2() < SIMD_EPSILON; + } + SIMD_FORCE_INLINE void serialize(struct btVector3Data& dataOut) const; + + SIMD_FORCE_INLINE void deSerialize(const struct btVector3Data& dataIn); + + SIMD_FORCE_INLINE void serializeFloat(struct btVector3FloatData& dataOut) const; + + SIMD_FORCE_INLINE void deSerializeFloat(const struct btVector3FloatData& dataIn); + + SIMD_FORCE_INLINE void serializeDouble(struct btVector3DoubleData& dataOut) const; + + SIMD_FORCE_INLINE void deSerializeDouble(const struct btVector3DoubleData& dataIn); + + /**@brief returns index of maximum dot product between this and vectors in array[] + * @param array The other vectors + * @param array_count The number of other vectors + * @param dotOut The maximum dot product */ + SIMD_FORCE_INLINE long maxDot( const btVector3 *array, long array_count, btScalar &dotOut ) const; + + /**@brief returns index of minimum dot product between this and vectors in array[] + * @param array The other vectors + * @param array_count The number of other vectors + * @param dotOut The minimum dot product */ + SIMD_FORCE_INLINE long minDot( const btVector3 *array, long array_count, btScalar &dotOut ) const; + + /* create a vector as btVector3( this->dot( btVector3 v0 ), this->dot( btVector3 v1), this->dot( btVector3 v2 )) */ + SIMD_FORCE_INLINE btVector3 dot3( const btVector3 &v0, const btVector3 &v1, const btVector3 &v2 ) const + { +#if defined BT_USE_SIMD_VECTOR3 && defined (BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + + __m128 a0 = _mm_mul_ps( v0.mVec128, this->mVec128 ); + __m128 a1 = _mm_mul_ps( v1.mVec128, this->mVec128 ); + __m128 a2 = _mm_mul_ps( v2.mVec128, this->mVec128 ); + __m128 b0 = _mm_unpacklo_ps( a0, a1 ); + __m128 b1 = _mm_unpackhi_ps( a0, a1 ); + __m128 b2 = _mm_unpacklo_ps( a2, _mm_setzero_ps() ); + __m128 r = _mm_movelh_ps( b0, b2 ); + r = _mm_add_ps( r, _mm_movehl_ps( b2, b0 )); + a2 = _mm_and_ps( a2, btvxyzMaskf); + r = _mm_add_ps( r, btCastdTo128f (_mm_move_sd( btCastfTo128d(a2), btCastfTo128d(b1) ))); + return btVector3(r); + +#elif defined(BT_USE_NEON) + static const uint32x4_t xyzMask = (const uint32x4_t){ -1, -1, -1, 0 }; + float32x4_t a0 = vmulq_f32( v0.mVec128, this->mVec128); + float32x4_t a1 = vmulq_f32( v1.mVec128, this->mVec128); + float32x4_t a2 = vmulq_f32( v2.mVec128, this->mVec128); + float32x2x2_t zLo = vtrn_f32( vget_high_f32(a0), vget_high_f32(a1)); + a2 = (float32x4_t) vandq_u32((uint32x4_t) a2, xyzMask ); + float32x2_t b0 = vadd_f32( vpadd_f32( vget_low_f32(a0), vget_low_f32(a1)), zLo.val[0] ); + float32x2_t b1 = vpadd_f32( vpadd_f32( vget_low_f32(a2), vget_high_f32(a2)), vdup_n_f32(0.0f)); + return btVector3( vcombine_f32(b0, b1) ); +#else + return btVector3( dot(v0), dot(v1), dot(v2)); +#endif + } }; /**@brief Return the sum of two vectors (Point symantics)*/ SIMD_FORCE_INLINE btVector3 operator+(const btVector3& v1, const btVector3& v2) { - return btVector3(v1.m_floats[0] + v2.m_floats[0], v1.m_floats[1] + v2.m_floats[1], v1.m_floats[2] + v2.m_floats[2]); +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + return btVector3(_mm_add_ps(v1.mVec128, v2.mVec128)); +#elif defined(BT_USE_NEON) + return btVector3(vaddq_f32(v1.mVec128, v2.mVec128)); +#else + return btVector3( + v1.m_floats[0] + v2.m_floats[0], + v1.m_floats[1] + v2.m_floats[1], + v1.m_floats[2] + v2.m_floats[2]); +#endif } /**@brief Return the elementwise product of two vectors */ SIMD_FORCE_INLINE btVector3 operator*(const btVector3& v1, const btVector3& v2) { - return btVector3(v1.m_floats[0] * v2.m_floats[0], v1.m_floats[1] * v2.m_floats[1], v1.m_floats[2] * v2.m_floats[2]); +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + return btVector3(_mm_mul_ps(v1.mVec128, v2.mVec128)); +#elif defined(BT_USE_NEON) + return btVector3(vmulq_f32(v1.mVec128, v2.mVec128)); +#else + return btVector3( + v1.m_floats[0] * v2.m_floats[0], + v1.m_floats[1] * v2.m_floats[1], + v1.m_floats[2] * v2.m_floats[2]); +#endif } /**@brief Return the difference between two vectors */ SIMD_FORCE_INLINE btVector3 operator-(const btVector3& v1, const btVector3& v2) { - return btVector3(v1.m_floats[0] - v2.m_floats[0], v1.m_floats[1] - v2.m_floats[1], v1.m_floats[2] - v2.m_floats[2]); +#if defined BT_USE_SIMD_VECTOR3 && (defined(BT_USE_SSE_IN_API) && defined(BT_USE_SSE)) + + // without _mm_and_ps this code causes slowdown in Concave moving + __m128 r = _mm_sub_ps(v1.mVec128, v2.mVec128); + return btVector3(_mm_and_ps(r, btvFFF0fMask)); +#elif defined(BT_USE_NEON) + float32x4_t r = vsubq_f32(v1.mVec128, v2.mVec128); + return btVector3((float32x4_t)vandq_s32((int32x4_t)r, btvFFF0Mask)); +#else + return btVector3( + v1.m_floats[0] - v2.m_floats[0], + v1.m_floats[1] - v2.m_floats[1], + v1.m_floats[2] - v2.m_floats[2]); +#endif } + /**@brief Return the negative of the vector */ SIMD_FORCE_INLINE btVector3 operator-(const btVector3& v) { +#if defined BT_USE_SIMD_VECTOR3 && (defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE)) + __m128 r = _mm_xor_ps(v.mVec128, btvMzeroMask); + return btVector3(_mm_and_ps(r, btvFFF0fMask)); +#elif defined(BT_USE_NEON) + return btVector3((btSimdFloat4)veorq_s32((int32x4_t)v.mVec128, (int32x4_t)btvMzeroMask)); +#else return btVector3(-v.m_floats[0], -v.m_floats[1], -v.m_floats[2]); +#endif } /**@brief Return the vector scaled by s */ SIMD_FORCE_INLINE btVector3 operator*(const btVector3& v, const btScalar& s) { +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + __m128 vs = _mm_load_ss(&s); // (S 0 0 0) + vs = bt_pshufd_ps(vs, 0x80); // (S S S 0.0) + return btVector3(_mm_mul_ps(v.mVec128, vs)); +#elif defined(BT_USE_NEON) + float32x4_t r = vmulq_n_f32(v.mVec128, s); + return btVector3((float32x4_t)vandq_s32((int32x4_t)r, btvFFF0Mask)); +#else return btVector3(v.m_floats[0] * s, v.m_floats[1] * s, v.m_floats[2] * s); +#endif } /**@brief Return the vector scaled by s */ @@ -411,14 +827,46 @@ SIMD_FORCE_INLINE btVector3 operator/(const btVector3& v, const btScalar& s) { btFullAssert(s != btScalar(0.0)); +#if 0 //defined(BT_USE_SSE_IN_API) +// this code is not faster ! + __m128 vs = _mm_load_ss(&s); + vs = _mm_div_ss(v1110, vs); + vs = bt_pshufd_ps(vs, 0x00); // (S S S S) + + return btVector3(_mm_mul_ps(v.mVec128, vs)); +#else return v * (btScalar(1.0) / s); +#endif } /**@brief Return the vector inversely scaled by s */ SIMD_FORCE_INLINE btVector3 operator/(const btVector3& v1, const btVector3& v2) { - return btVector3(v1.m_floats[0] / v2.m_floats[0],v1.m_floats[1] / v2.m_floats[1],v1.m_floats[2] / v2.m_floats[2]); +#if defined BT_USE_SIMD_VECTOR3 && (defined(BT_USE_SSE_IN_API)&& defined (BT_USE_SSE)) + __m128 vec = _mm_div_ps(v1.mVec128, v2.mVec128); + vec = _mm_and_ps(vec, btvFFF0fMask); + return btVector3(vec); +#elif defined(BT_USE_NEON) + float32x4_t x, y, v, m; + + x = v1.mVec128; + y = v2.mVec128; + + v = vrecpeq_f32(y); // v ~ 1/y + m = vrecpsq_f32(y, v); // m = (2-v*y) + v = vmulq_f32(v, m); // vv = v*m ~~ 1/y + m = vrecpsq_f32(y, v); // mm = (2-vv*y) + v = vmulq_f32(v, x); // x*vv + v = vmulq_f32(v, m); // (x*vv)*(2-vv*y) = x*(vv(2-vv*y)) ~~~ x/y + + return btVector3(v); +#else + return btVector3( + v1.m_floats[0] / v2.m_floats[0], + v1.m_floats[1] / v2.m_floats[1], + v1.m_floats[2] / v2.m_floats[2]); +#endif } /**@brief Return the dot product between two vectors */ @@ -488,22 +936,133 @@ SIMD_FORCE_INLINE btScalar btVector3::distance(const btVector3& v) const SIMD_FORCE_INLINE btVector3 btVector3::normalized() const { +#if defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + btVector3 norm = *this; + + return norm.normalize(); +#else return *this / length(); +#endif } -SIMD_FORCE_INLINE btVector3 btVector3::rotate( const btVector3& wAxis, const btScalar angle ) const +SIMD_FORCE_INLINE btVector3 btVector3::rotate( const btVector3& wAxis, const btScalar _angle ) const { // wAxis must be a unit lenght vector +#if defined BT_USE_SIMD_VECTOR3 && defined (BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + + __m128 O = _mm_mul_ps(wAxis.mVec128, mVec128); + btScalar ssin = btSin( _angle ); + __m128 C = wAxis.cross( mVec128 ).mVec128; + O = _mm_and_ps(O, btvFFF0fMask); + btScalar scos = btCos( _angle ); + + __m128 vsin = _mm_load_ss(&ssin); // (S 0 0 0) + __m128 vcos = _mm_load_ss(&scos); // (S 0 0 0) + + __m128 Y = bt_pshufd_ps(O, 0xC9); // (Y Z X 0) + __m128 Z = bt_pshufd_ps(O, 0xD2); // (Z X Y 0) + O = _mm_add_ps(O, Y); + vsin = bt_pshufd_ps(vsin, 0x80); // (S S S 0) + O = _mm_add_ps(O, Z); + vcos = bt_pshufd_ps(vcos, 0x80); // (S S S 0) + + vsin = vsin * C; + O = O * wAxis.mVec128; + __m128 X = mVec128 - O; + + O = O + vsin; + vcos = vcos * X; + O = O + vcos; + + return btVector3(O); +#else btVector3 o = wAxis * wAxis.dot( *this ); - btVector3 x = *this - o; - btVector3 y; + btVector3 _x = *this - o; + btVector3 _y; - y = wAxis.cross( *this ); + _y = wAxis.cross( *this ); - return ( o + x * btCos( angle ) + y * btSin( angle ) ); + return ( o + _x * btCos( _angle ) + _y * btSin( _angle ) ); +#endif +} + +SIMD_FORCE_INLINE long btVector3::maxDot( const btVector3 *array, long array_count, btScalar &dotOut ) const +{ +#if (defined BT_USE_SSE && defined BT_USE_SIMD_VECTOR3 && defined BT_USE_SSE_IN_API) || defined (BT_USE_NEON) + #if defined _WIN32 || defined (BT_USE_SSE) + const long scalar_cutoff = 10; + long _maxdot_large( const float *array, const float *vec, unsigned long array_count, float *dotOut ); + #elif defined BT_USE_NEON + const long scalar_cutoff = 4; + extern long (*_maxdot_large)( const float *array, const float *vec, unsigned long array_count, float *dotOut ); + #endif + if( array_count < scalar_cutoff ) +#endif + { + btScalar maxDot = -SIMD_INFINITY; + int i = 0; + int ptIndex = -1; + for( i = 0; i < array_count; i++ ) + { + btScalar dot = array[i].dot(*this); + + if( dot > maxDot ) + { + maxDot = dot; + ptIndex = i; + } + } + + dotOut = maxDot; + return ptIndex; + } +#if (defined BT_USE_SSE && defined BT_USE_SIMD_VECTOR3 && defined BT_USE_SSE_IN_API) || defined (BT_USE_NEON) + return _maxdot_large( (float*) array, (float*) &m_floats[0], array_count, &dotOut ); +#endif } +SIMD_FORCE_INLINE long btVector3::minDot( const btVector3 *array, long array_count, btScalar &dotOut ) const +{ +#if (defined BT_USE_SSE && defined BT_USE_SIMD_VECTOR3 && defined BT_USE_SSE_IN_API) || defined (BT_USE_NEON) + #if defined BT_USE_SSE + const long scalar_cutoff = 10; + long _mindot_large( const float *array, const float *vec, unsigned long array_count, float *dotOut ); + #elif defined BT_USE_NEON + const long scalar_cutoff = 4; + extern long (*_mindot_large)( const float *array, const float *vec, unsigned long array_count, float *dotOut ); + #else + #error unhandled arch! + #endif + + if( array_count < scalar_cutoff ) +#endif + { + btScalar minDot = SIMD_INFINITY; + int i = 0; + int ptIndex = -1; + + for( i = 0; i < array_count; i++ ) + { + btScalar dot = array[i].dot(*this); + + if( dot < minDot ) + { + minDot = dot; + ptIndex = i; + } + } + + dotOut = minDot; + + return ptIndex; + } +#if (defined BT_USE_SSE && defined BT_USE_SIMD_VECTOR3 && defined BT_USE_SSE_IN_API) || defined (BT_USE_NEON) + return _mindot_large( (float*) array, (float*) &m_floats[0], array_count, &dotOut ); +#endif//BT_USE_SIMD_VECTOR3 +} + + class btVector4 : public btVector3 { public: @@ -511,24 +1070,47 @@ public: SIMD_FORCE_INLINE btVector4() {} - SIMD_FORCE_INLINE btVector4(const btScalar& x, const btScalar& y, const btScalar& z,const btScalar& w) - : btVector3(x,y,z) + SIMD_FORCE_INLINE btVector4(const btScalar& _x, const btScalar& _y, const btScalar& _z,const btScalar& _w) + : btVector3(_x,_y,_z) + { + m_floats[3] = _w; + } + +#if (defined (BT_USE_SSE_IN_API)&& defined (BT_USE_SSE)) || defined (BT_USE_NEON) + SIMD_FORCE_INLINE btVector4(const btSimdFloat4 vec) + { + mVec128 = vec; + } + + SIMD_FORCE_INLINE btVector4(const btVector3& rhs) { - m_floats[3] = w; + mVec128 = rhs.mVec128; } + SIMD_FORCE_INLINE btVector4& + operator=(const btVector4& v) + { + mVec128 = v.mVec128; + return *this; + } +#endif // #if defined (BT_USE_SSE_IN_API) || defined (BT_USE_NEON) SIMD_FORCE_INLINE btVector4 absolute4() const { +#if defined BT_USE_SIMD_VECTOR3 && defined(BT_USE_SSE_IN_API) && defined (BT_USE_SSE) + return btVector4(_mm_and_ps(mVec128, btvAbsfMask)); +#elif defined(BT_USE_NEON) + return btVector4(vabsq_f32(mVec128)); +#else return btVector4( btFabs(m_floats[0]), btFabs(m_floats[1]), btFabs(m_floats[2]), btFabs(m_floats[3])); +#endif } - btScalar getW() const { return m_floats[3];} @@ -556,12 +1138,8 @@ public: maxIndex = 3; maxVal = m_floats[3]; } - - - return maxIndex; - } @@ -591,7 +1169,6 @@ public: } return minIndex; - } @@ -623,12 +1200,12 @@ public: * @param z Value of z * @param w Value of w */ - SIMD_FORCE_INLINE void setValue(const btScalar& x, const btScalar& y, const btScalar& z,const btScalar& w) + SIMD_FORCE_INLINE void setValue(const btScalar& _x, const btScalar& _y, const btScalar& _z,const btScalar& _w) { - m_floats[0]=x; - m_floats[1]=y; - m_floats[2]=z; - m_floats[3]=w; + m_floats[0]=_x; + m_floats[1]=_y; + m_floats[2]=_z; + m_floats[3]=_w; } @@ -762,5 +1339,4 @@ SIMD_FORCE_INLINE void btVector3::deSerialize(const struct btVector3Data& dataIn m_floats[i] = dataIn.m_floats[i]; } - #endif //BT_VECTOR3_H |