Skip to content

O
opencv
Commit 283348af authored by Vitaly Tuzov's avatar Vitaly Tuzov
Browse files
Options

SSE2 code in invert() replaced with universal intrinsics

parent 76d4aa0c
Hide whitespace changes
Inline Side-by-side
Showing with 36 additions and 64 deletions
modules/core/src/lapack.cpp
View file @ 283348af
...@@ -910,37 +910,23 @@ double cv::invert( InputArray _src, OutputArray _dst, int method ) ...@@ -910,37 +910,23 @@ double cv::invert( InputArray _src, OutputArray _dst, int method )
{ {
result = true; result = true;
d = 1./d; d = 1./d;
#if CV_SIMD128
#if CV_SSE2 static const float CV_DECL_ALIGNED(16) inv[4] = { 0.f,-0.f,-0.f,0.f };
if(USE_SSE2) v_float32x4 s0 = (v_load_halves((const float*)srcdata, (const float*)(srcdata + srcstep)) * v_setall_f32((float)d)) ^ v_load((const float *)inv);//0123//3120
{ s0 = v_extract<3>(s0, v_combine_low(v_rotate_right<1>(s0), s0));
__m128 zero = _mm_setzero_ps(); v_store_low((float*)dstdata, s0);
__m128 t0 = _mm_loadl_pi(zero, (const __m64*)srcdata); //t0 = sf(0,0) sf(0,1) v_store_high((float*)(dstdata + dststep), s0);
__m128 t1 = _mm_loadh_pi(zero, (const __m64*)(srcdata+srcstep)); //t1 = sf(1,0) sf(1,1) #else
__m128 s0 = _mm_or_ps(t0, t1); double t0, t1;
__m128 det =_mm_set1_ps((float)d); t0 = Sf(0,0)*d;
s0 = _mm_mul_ps(s0, det); t1 = Sf(1,1)*d;
static const uchar CV_DECL_ALIGNED(16) inv[16] = {0,0,0,0,0,0,0,0x80,0,0,0,0x80,0,0,0,0}; Df(1,1) = (float)t0;
__m128 pattern = _mm_load_ps((const float*)inv); Df(0,0) = (float)t1;
s0 = _mm_xor_ps(s0, pattern);//==-1*s0 t0 = -Sf(0,1)*d;
s0 = _mm_shuffle_ps(s0, s0, _MM_SHUFFLE(0,2,1,3)); t1 = -Sf(1,0)*d;
_mm_storel_pi((__m64*)dstdata, s0); Df(0,1) = (float)t0;
_mm_storeh_pi((__m64*)((float*)(dstdata+dststep)), s0); Df(1,0) = (float)t1;
} #endif
else
#endif
{
double t0, t1;
t0 = Sf(0,0)*d;
t1 = Sf(1,1)*d;
Df(1,1) = (float)t0;
Df(0,0) = (float)t1;
t0 = -Sf(0,1)*d;
t1 = -Sf(1,0)*d;
Df(0,1) = (float)t0;
Df(1,0) = (float)t1;
}
} }
} }
else else
...@@ -950,39 +936,25 @@ double cv::invert( InputArray _src, OutputArray _dst, int method ) ...@@ -950,39 +936,25 @@ double cv::invert( InputArray _src, OutputArray _dst, int method )
{ {
result = true; result = true;
d = 1./d; d = 1./d;
#if CV_SSE2 #if CV_SIMD128_64F
if(USE_SSE2) v_float64x2 det = v_setall_f64(d);
{ v_float64x2 s0 = v_load((const double*)srcdata) * det;
__m128d s0 = _mm_loadu_pd((const double*)srcdata); //s0 = sf(0,0) sf(0,1) v_float64x2 s1 = v_load((const double*)(srcdata+srcstep)) * det;
__m128d s1 = _mm_loadu_pd ((const double*)(srcdata+srcstep));//s1 = sf(1,0) sf(1,1) v_float64x2 sm = v_extract<1>(s1, s0);//30
__m128d sm = _mm_unpacklo_pd(s0, _mm_load_sd((const double*)(srcdata+srcstep)+1)); //sm = sf(0,0) sf(1,1) - main diagonal v_float64x2 ss = v_extract<1>(s0, s1) ^ v_setall_f64(-0.);//12
__m128d ss = _mm_shuffle_pd(s0, s1, _MM_SHUFFLE2(0,1)); //ss = sf(0,1) sf(1,0) - secondary diagonal v_store((double*)dstdata, v_combine_low(sm, ss));//31
__m128d det = _mm_load1_pd((const double*)&d); v_store((double*)(dstdata + dststep), v_combine_high(ss, sm));//20
sm = _mm_mul_pd(sm, det); #else
double t0, t1;
static const uchar CV_DECL_ALIGNED(16) inv[8] = {0,0,0,0,0,0,0,0x80}; t0 = Sd(0,0)*d;
__m128d pattern = _mm_load1_pd((double*)inv); t1 = Sd(1,1)*d;
ss = _mm_mul_pd(ss, det); Dd(1,1) = t0;
ss = _mm_xor_pd(ss, pattern);//==-1*ss Dd(0,0) = t1;
t0 = -Sd(0,1)*d;
s0 = _mm_shuffle_pd(sm, ss, _MM_SHUFFLE2(0,1)); t1 = -Sd(1,0)*d;
s1 = _mm_shuffle_pd(ss, sm, _MM_SHUFFLE2(0,1)); Dd(0,1) = t0;
_mm_storeu_pd((double*)dstdata, s0); Dd(1,0) = t1;
_mm_storeu_pd((double*)(dstdata+dststep), s1); #endif
}
else
#endif
{
double t0, t1;
t0 = Sd(0,0)*d;
t1 = Sd(1,1)*d;
Dd(1,1) = t0;
Dd(0,0) = t1;
t0 = -Sd(0,1)*d;
t1 = -Sd(1,0)*d;
Dd(0,1) = t0;
Dd(1,0) = t1;
}
} }
} }
} }
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment