third_party/opencv
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fixed v_signmask implementation for AVX2, updated universal intrinsics tests.

Browse Source
This commit is contained in:
Vitaly Tuzov
2019-05-23 13:45:14 +03:00
parent def8fa2248
commit f0fb91f2d4
2 changed files with 7 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/core/include/opencv2/core/hal/intrin_avx.hpp
+3 -7
View File
@@ -1227,18 +1227,14 @@ inline int v_signmask(const v_uint8x32& a)
{ return v_signmask(v_reinterpret_as_s8(a)); }
inline int v_signmask(const v_int16x16& a)
{
v_int8x32 v = v_int8x32(_mm256_packs_epi16(a.val, a.val));
return v_signmask(v) & 255;
}
{ return v_signmask(v_pack(a, a)) & 0xFFFF; }
inline int v_signmask(const v_uint16x16& a)
{ return v_signmask(v_reinterpret_as_s16(a)); }
inline int v_signmask(const v_int32x8& a)
{
__m256i a16 = _mm256_packs_epi32(a.val, a.val);
v_int8x32 v = v_int8x32(_mm256_packs_epi16(a16, a16));
return v_signmask(v) & 15;
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)); }