Merge pull request #10136 from alalek:issue_10134

pull/10076/merge
Alexander Alekhin 7 years ago
commit e4aa2ccd66
  1. 16
      modules/core/include/opencv2/core/hal/intrin_sse.hpp
  2. 12
      modules/core/include/opencv2/core/hal/intrin_vsx.hpp

@ -1024,24 +1024,28 @@ OPENCV_HAL_IMPL_SSE_SHIFT_OP(v_uint64x2, v_int64x2, epi64, v_srai_epi64)
template<int imm, typename _Tpvec> template<int imm, typename _Tpvec>
inline _Tpvec v_rotate_right(const _Tpvec &a) inline _Tpvec v_rotate_right(const _Tpvec &a)
{ {
return _Tpvec(_mm_srli_si128(a.val, imm*(sizeof(typename _Tpvec::lane_type)))); enum { CV_SHIFT = imm*(sizeof(typename _Tpvec::lane_type)) };
return _Tpvec(_mm_srli_si128(a.val, CV_SHIFT));
} }
template<int imm, typename _Tpvec> template<int imm, typename _Tpvec>
inline _Tpvec v_rotate_left(const _Tpvec &a) inline _Tpvec v_rotate_left(const _Tpvec &a)
{ {
return _Tpvec(_mm_slli_si128(a.val, imm*(sizeof(typename _Tpvec::lane_type)))); enum { CV_SHIFT = imm*(sizeof(typename _Tpvec::lane_type)) };
return _Tpvec(_mm_slli_si128(a.val, CV_SHIFT));
} }
template<int imm, typename _Tpvec> template<int imm, typename _Tpvec>
inline _Tpvec v_rotate_right(const _Tpvec &a, const _Tpvec &b) inline _Tpvec v_rotate_right(const _Tpvec &a, const _Tpvec &b)
{ {
const int cWidth = sizeof(typename _Tpvec::lane_type); enum { CV_SHIFT1 = imm*(sizeof(typename _Tpvec::lane_type)) };
return _Tpvec(_mm_or_si128(_mm_srli_si128(a.val, imm*cWidth), _mm_slli_si128(b.val, (16 - imm*cWidth)))); enum { CV_SHIFT2 = 16 - imm*(sizeof(typename _Tpvec::lane_type)) };
return _Tpvec(_mm_or_si128(_mm_srli_si128(a.val, CV_SHIFT1), _mm_slli_si128(b.val, CV_SHIFT2)));
} }
template<int imm, typename _Tpvec> template<int imm, typename _Tpvec>
inline _Tpvec v_rotate_left(const _Tpvec &a, const _Tpvec &b) inline _Tpvec v_rotate_left(const _Tpvec &a, const _Tpvec &b)
{ {
const int cWidth = sizeof(typename _Tpvec::lane_type); enum { CV_SHIFT1 = imm*(sizeof(typename _Tpvec::lane_type)) };
return _Tpvec(_mm_or_si128(_mm_slli_si128(a.val, imm*cWidth), _mm_srli_si128(b.val, (16 - imm*cWidth)))); enum { CV_SHIFT2 = 16 - imm*(sizeof(typename _Tpvec::lane_type)) };
return _Tpvec(_mm_or_si128(_mm_slli_si128(a.val, CV_SHIFT1), _mm_srli_si128(b.val, CV_SHIFT2)));
} }
#define OPENCV_HAL_IMPL_SSE_LOADSTORE_INT_OP(_Tpvec, _Tp) \ #define OPENCV_HAL_IMPL_SSE_LOADSTORE_INT_OP(_Tpvec, _Tp) \

@ -634,19 +634,19 @@ OPENCV_IMPL_VSX_ROTATE_LR(v_int64x2, vec_dword2)
template<int imm, typename _Tpvec> template<int imm, typename _Tpvec>
inline _Tpvec v_rotate_right(const _Tpvec& a, const _Tpvec& b) inline _Tpvec v_rotate_right(const _Tpvec& a, const _Tpvec& b)
{ {
const int wd = imm * sizeof(typename _Tpvec::lane_type); enum { CV_SHIFT = 16 - imm * (sizeof(typename _Tpvec::lane_type)) };
if (wd == 0) if (CV_SHIFT == 16)
return a; return a;
return _Tpvec(vec_sld(b.val, a.val, 16 - wd)); return _Tpvec(vec_sld(b.val, a.val, CV_SHIFT));
} }
template<int imm, typename _Tpvec> template<int imm, typename _Tpvec>
inline _Tpvec v_rotate_left(const _Tpvec& a, const _Tpvec& b) inline _Tpvec v_rotate_left(const _Tpvec& a, const _Tpvec& b)
{ {
const int wd = imm * sizeof(typename _Tpvec::lane_type); enum { CV_SHIFT = imm * (sizeof(typename _Tpvec::lane_type)) };
if (wd == 16) if (CV_SHIFT == 16)
return b; return b;
return _Tpvec(vec_sld(a.val, b.val, wd)); return _Tpvec(vec_sld(a.val, b.val, CV_SHIFT));
} }
#define OPENCV_IMPL_VSX_ROTATE_64(_Tpvec, suffix, rg1, rg2) \ #define OPENCV_IMPL_VSX_ROTATE_64(_Tpvec, suffix, rg1, rg2) \

Loading…
Cancel
Save