mirror/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2025-06-10 11:03:03 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Fix for AVX2 implementation of v_check_any(), v_check_all() intrinsics

Browse Source
This commit is contained in:
Vitaly Tuzov 2019-08-27 14:31:23 +03:00
parent 824465ea27
commit 1b40528e1a
1 changed files with 9 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
modules/core/include/opencv2/core/hal/intrin_avx.hpp
View File

@ -1231,19 +1231,16 @@ inline int v_signmask(const v_int16x16& a)
inline int v_signmask(const v_uint16x16& a) inline int v_signmask(const v_uint16x16& a)
{ return v_signmask(v_reinterpret_as_s16(a)); } { return v_signmask(v_reinterpret_as_s16(a)); }
inline int v_signmask(const v_int32x8& a)
{
v_int16x16 a16 = v_pack(a, a);
return v_signmask(v_pack(a16, a16)) & 0xFF;
}
inline int v_signmask(const v_uint32x8& a)
{ return v_signmask(v_reinterpret_as_s32(a)); }
inline int v_signmask(const v_float32x8& a) inline int v_signmask(const v_float32x8& a)
{ return _mm256_movemask_ps(a.val); } { return _mm256_movemask_ps(a.val); }
inline int v_signmask(const v_float64x4& a) inline int v_signmask(const v_float64x4& a)
{ return _mm256_movemask_pd(a.val); } { return _mm256_movemask_pd(a.val); }
inline int v_signmask(const v_int32x8& a)
{ return v_signmask(v_reinterpret_as_f32(a)); }
inline int v_signmask(const v_uint32x8& a)
{ return v_signmask(v_reinterpret_as_f32(a)); }
inline int v_scan_forward(const v_int8x32& a) { return trailingZeros32(v_signmask(v_reinterpret_as_s8(a))); } inline int v_scan_forward(const v_int8x32& a) { return trailingZeros32(v_signmask(v_reinterpret_as_s8(a))); }
inline int v_scan_forward(const v_uint8x32& a) { return trailingZeros32(v_signmask(v_reinterpret_as_s8(a))); } inline int v_scan_forward(const v_uint8x32& a) { return trailingZeros32(v_signmask(v_reinterpret_as_s8(a))); }
inline int v_scan_forward(const v_int16x16& a) { return trailingZeros32(v_signmask(v_reinterpret_as_s8(a))) / 2; } inline int v_scan_forward(const v_int16x16& a) { return trailingZeros32(v_signmask(v_reinterpret_as_s8(a))) / 2; }
@ -1270,10 +1267,10 @@ inline int v_scan_forward(const v_float64x4& a) { return trailingZeros32(v_signm
OPENCV_HAL_IMPL_AVX_CHECK(v_uint8x32, OPENCV_HAL_1ST, -1) OPENCV_HAL_IMPL_AVX_CHECK(v_uint8x32, OPENCV_HAL_1ST, -1)
OPENCV_HAL_IMPL_AVX_CHECK(v_int8x32, OPENCV_HAL_1ST, -1) OPENCV_HAL_IMPL_AVX_CHECK(v_int8x32, OPENCV_HAL_1ST, -1)
OPENCV_HAL_IMPL_AVX_CHECK(v_uint16x16, OPENCV_HAL_AND, (int)0xaaaa) OPENCV_HAL_IMPL_AVX_CHECK(v_uint16x16, OPENCV_HAL_AND, (int)0xaaaaaaaa)
OPENCV_HAL_IMPL_AVX_CHECK(v_int16x16, OPENCV_HAL_AND, (int)0xaaaa) OPENCV_HAL_IMPL_AVX_CHECK(v_int16x16, OPENCV_HAL_AND, (int)0xaaaaaaaa)
OPENCV_HAL_IMPL_AVX_CHECK(v_uint32x8, OPENCV_HAL_AND, (int)0x8888) OPENCV_HAL_IMPL_AVX_CHECK(v_uint32x8, OPENCV_HAL_AND, (int)0x88888888)
OPENCV_HAL_IMPL_AVX_CHECK(v_int32x8, OPENCV_HAL_AND, (int)0x8888) OPENCV_HAL_IMPL_AVX_CHECK(v_int32x8, OPENCV_HAL_AND, (int)0x88888888)
#define OPENCV_HAL_IMPL_AVX_CHECK_FLT(_Tpvec, allmask) \ #define OPENCV_HAL_IMPL_AVX_CHECK_FLT(_Tpvec, allmask) \
inline bool v_check_all(const _Tpvec& a) \ inline bool v_check_all(const _Tpvec& a) \