Added SSE det function

This commit is contained in:
Christophe Riccio 2011-01-06 18:33:42 +00:00
parent e3ae7e3db5
commit a9cef7a516
3 changed files with 76 additions and 37 deletions

View File

@ -181,7 +181,7 @@ inline void _mm_transpose_ps(__m128 const in[4], __m128 out[4])
out[3] = _mm_shuffle_ps(tmp2, tmp3, 0xDD); out[3] = _mm_shuffle_ps(tmp2, tmp3, 0xDD);
} }
inline __m128 _mm_det_ps(__m128 const in[4]) inline __m128 _mm_slow_det_ps(__m128 const in[4])
{ {
__m128 Fac0; __m128 Fac0;
{ {
@ -407,14 +407,29 @@ inline __m128 _mm_det_ps(__m128 const in[4])
return Det0; return Det0;
} }
template <typename T> /*
inline typename detail::tmat4x4<T>::value_type _mm_det2_ps float ssedot(const float * __restrict u, const float * __restrict v){
( __m128 uv = _mm_mul_ps(_mm_load_ps(u), _mm_load_ps(v));
__m128 const & m[4] uv = _mm_hadd_ps(uv, uv); // or shuffle like there's no tomorrow
) uv = _mm_hadd_ps(uv, uv); // if there ain't no haddps around.
{ return __builtin_ia32_vec_ext_v4sf(uv, 0);
GLM_STATIC_ASSERT(detail::type<T>::is_float, "'determinant' only accept floating-point inputs"); }
float dot_sse(float *u, float *v)
{
__m128 mul = _mm_mul_ps(_mm_load_ps(u), _mm_load_ps(v)); // { a,b,c,d }
__m128 hi = _mm_movehl_ps(mul, mul); // { c,d,c,d }
__m128 add = _mm_add_ps(mul, hi); // { a+c, b+d, c+c, d+d }
__m128 half = _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(add), _MM_SHUFFLE(1, 1, 1, 1))); // { b+d }
__m128 join = _mm_add_ss(add, half); // { a+c+b+d }
float f;
_mm_store_ss(&f, join);
return f;
}
*/
inline __m128 _mm_det_ps(__m128 const m[4])
{
//T SubFactor00 = m[2][2] * m[3][3] - m[3][2] * m[2][3]; //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 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 SubFactor02 = m[2][1] * m[3][2] - m[3][1] * m[2][2];
@ -422,24 +437,27 @@ inline typename detail::tmat4x4<T>::value_type _mm_det2_ps
//T SubFactor04 = m[2][0] * m[3][2] - m[3][0] * m[2][2]; //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]; //T SubFactor05 = m[2][0] * m[3][1] - m[3][0] * m[2][1];
__m128 const & m2 = m[2];
__m128 const & m3 = m[3];
// First 2 columns // First 2 columns
__m128 Swp2A = _mm_shuffle_ps(m[2], m[2], _MM_SHUFFLE(0, 1, 1, 2)); __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 Swp3A = _mm_shuffle_ps(m[3], m[3], _MM_SHUFFLE(3, 2, 3, 3));
__m128 MulA = __mm_mul_ps(Swp2A, Swp3A); __m128 MulA = _mm_mul_ps(Swp2A, Swp3A);
// Second 2 columns // Second 2 columns
__m128 Swp2B = _mm_shuffle_ps(m[2], m[2], _MM_SHUFFLE(0, 1, 1, 2)); __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(3, 2, 3, 3)); __m128 Swp3B = _mm_shuffle_ps(m[3], m[3], _MM_SHUFFLE(0, 1, 1, 2));
__m128 MulB = __mm_mul_ps(Swp2A, Swp3A); __m128 MulB = _mm_mul_ps(Swp2B, Swp3B);
// Columns subtraction // Columns subtraction
__m128 SubAB = __mm_sub_ps(MulA, MulB); __m128 SubE = _mm_sub_ps(MulA, MulB);
// Last 2 rows // Last 2 rows
__m128 Swp2C = _mm_shuffle_ps(m[2], m[2], _MM_SHUFFLE(1, 2, 0, 0)); __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(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 MulC = _mm_mul_ps(Swp2C, Swp3C);
__m128 SwpD = __mm_hl_ps(MulC); __m128 SubF = _mm_sub_ps(_mm_movehl_ps(MulC, MulC), MulC);
//detail::tvec4<T> DetCof( //detail::tvec4<T> DetCof(
// + (m[1][1] * SubFactor00 - m[1][2] * SubFactor01 + m[1][3] * SubFactor02), // + (m[1][1] * SubFactor00 - m[1][2] * SubFactor01 + m[1][3] * SubFactor02),
@ -447,29 +465,31 @@ inline typename detail::tmat4x4<T>::value_type _mm_det2_ps
// + (m[1][0] * SubFactor01 - m[1][1] * SubFactor03 + m[1][3] * SubFactor05), // + (m[1][0] * SubFactor01 - m[1][1] * SubFactor03 + m[1][3] * SubFactor05),
// - (m[1][0] * SubFactor02 - m[1][1] * SubFactor04 + m[1][2] * SubFactor05)); // - (m[1][0] * SubFactor02 - m[1][1] * SubFactor04 + m[1][2] * SubFactor05));
__128 SubFacA = _mm_shuffle_ps(SubAB, SubAB, _MM_SHUFFLE(2, 1, 0, 0)); __m128 SubFacA = _mm_shuffle_ps(SubE, SubE, _MM_SHUFFLE(2, 1, 0, 0));
__128 SwpFacA = _mm_shuffle_ps(m[1], m[1], _MM_SHUFFLE(0, 0, 0, 1)); __m128 SwpFacA = _mm_shuffle_ps(m[1], m[1], _MM_SHUFFLE(0, 0, 0, 1));
__128 MulFacA = __mm_mul_ps(SwpFacA, SubFacA); __m128 MulFacA = _mm_mul_ps(SwpFacA, SubFacA);
__128 SubFacB = ; __m128 SubTmpB = _mm_shuffle_ps(SubE, SubF, _MM_SHUFFLE(0, 0, 3, 1));
__128 SwpFacB = _mm_shuffle_ps(m[1], m[1], _MM_SHUFFLE(1, 1, 2, 2)); __m128 SubFacB = _mm_shuffle_ps(SubTmpB, SubTmpB, _MM_SHUFFLE(3, 1, 1, 0));//SubF[0], SubE[3], SubE[3], SubE[1];
__128 MulFacB = __mm_mul_ps(SwpFacB, SubFacB); __m128 SwpFacB = _mm_shuffle_ps(m[1], m[1], _MM_SHUFFLE(1, 1, 2, 2));
__m128 MulFacB = _mm_mul_ps(SwpFacB, SubFacB);
__128 SubRes = __mm_sub_ps(MulFacA, MulFacA); __m128 SubRes = _mm_sub_ps(MulFacA, MulFacB);
__128 SubFacC = ; __m128 SubTmpC = _mm_shuffle_ps(SubE, SubF, _MM_SHUFFLE(1, 0, 2, 2));
__128 SwpFacC = _mm_shuffle_ps(m[1], m[1], _MM_SHUFFLE(1, 1, 2, 2)); __m128 SubFacC = _mm_shuffle_ps(SubTmpC, SubTmpC, _MM_SHUFFLE(3, 3, 2, 0));
__128 MulFacC = __mm_mul_ps(SwpFacC, SubFacC); __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 AddRes = _mm_add_ps(SubRes, MulFacC);
__m128 DetCof = __mm_mul_ps(AddRes, _mm_set_ps( 1.0f,-1.0f, 1.0f,-1.0f)); __m128 DetCof = _mm_mul_ps(AddRes, _mm_setr_ps( 1.0f,-1.0f, 1.0f,-1.0f));
//return m[0][0] * DetCof[0] //return m[0][0] * DetCof[0]
// + m[0][1] * DetCof[1] // + m[0][1] * DetCof[1]
// + m[0][2] * DetCof[2] // + m[0][2] * DetCof[2]
// + m[0][3] * DetCof[3]; // + m[0][3] * DetCof[3];
return _mm_dot_ps(m[0], Signed); return _mm_dot_ps(m[0], DetCof);
} }
inline void _mm_inverse_ps(__m128 const in[4], __m128 out[4]) inline void _mm_inverse_ps(__m128 const in[4], __m128 out[4])
@ -937,7 +957,6 @@ inline void _mm_inverse_fast_ps(__m128 const in[4], __m128 out[4])
out[3] = _mm_mul_ps(Inv3, Rcp0); out[3] = _mm_mul_ps(Inv3, Rcp0);
} }
void _mm_rotate_ps(__m128 const in[4], float Angle, float const v[3], __m128 out[4]) void _mm_rotate_ps(__m128 const in[4], float Angle, float const v[3], __m128 out[4])
{ {
float a = glm::radians(Angle); float a = glm::radians(Angle);
@ -1010,8 +1029,8 @@ void _mm_rotate_ps(__m128 const in[4], float Angle, float const v[3], __m128 out
void _mm_outer_ps(__m128 const & c, __m128 const & r, __m128 out[4]) void _mm_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[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[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[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)); out[3] = _mm_mul_ps(c, _mm_shuffle_ps(r, r, _MM_SHUFFLE(3, 3, 3, 3)));
} }

View File

@ -228,7 +228,7 @@ namespace simd_mat4
detail::fmat4x4SIMD const & y detail::fmat4x4SIMD const & y
) )
{ {
GLM_STATIC_ASSERT(0, "TODO"); //GLM_STATIC_ASSERT(0, "TODO");
} }
inline detail::fmat4x4SIMD simd_transpose(detail::fmat4x4SIMD const & m) inline detail::fmat4x4SIMD simd_transpose(detail::fmat4x4SIMD const & m)
@ -240,7 +240,7 @@ namespace simd_mat4
inline float simd_determinant(detail::fmat4x4SIMD const & m) inline float simd_determinant(detail::fmat4x4SIMD const & m)
{ {
GLM_STATIC_ASSERT(0, "TODO"); //GLM_STATIC_ASSERT(0, "TODO");
} }
inline detail::fmat4x4SIMD simd_inverse(detail::fmat4x4SIMD const & m) inline detail::fmat4x4SIMD simd_inverse(detail::fmat4x4SIMD const & m)

View File

@ -13,5 +13,25 @@
int main(int argc, void* argv[]) int main(int argc, void* argv[])
{ {
// shuffle test
glm::simd_vec4 A(1.0f, 2.0f, 3.0f, 4.0f);
glm::simd_vec4 B(5.0f, 6.0f, 7.0f, 8.0f);
__m128 C = _mm_shuffle_ps(A.Data, B.Data, _MM_SHUFFLE(1, 0, 1, 0));
glm::mat4 IdentityA(
glm::vec4(4.0f, 0.7f, 0.1f, 0.01f),
glm::vec4(0.5f, 3.0f, 0.6f, 0.02f),
glm::vec4(0.2f, 0.4f, 2.0f, 0.03f),
glm::vec4(4.0f, 3.0f, 2.0f, 1.00f));
float DetA = glm::determinant(IdentityA);
glm::simd_mat4 IdentityB(
glm::simd_vec4(4.0f, 0.7f, 0.1f, 0.01f),
glm::simd_vec4(0.5f, 3.0f, 0.6f, 0.02f),
glm::simd_vec4(0.2f, 0.4f, 2.0f, 0.03f),
glm::simd_vec4(4.0f, 3.0f, 2.0f, 1.00f));
__m128 DetB = _mm_slow_det_ps(&IdentityB.Data[0].Data);
__m128 DetC = _mm_det_ps(&IdentityB.Data[0].Data);
return 0;
} }