potok/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2026-07-31 00:03:03 +04:00
Code Packages Releases Wiki Activity

Change behaviour of 16-bit multiply operator

Browse Source 
- redefine 16-bit multiply operator to perform saturating multiply
    instead of non-saturating multiply
  - implement 8-bit multiply operator to perform saturating multiply
  - implement v_mul_wrap() for 8-bit, 16-bit non-saturating multiply
  - improve performance of v_mul_hi() for VSX
  - update intrin tests with new changes
  - replace unv 16-bit multiplication operator with v_mul_wrap due behavior changes

  - Several improvements depend on vpisarev review

    * initial forward declarations for universal intrinsics
    * move emulating SSE intrinsics into separate file
    * implement v_mul_expand for 8-bit
    * reimplement saturating multiply using v_mul_expand + v_pack
    * map v_expand, v_load_expand, v_load_expand_q to sse4.1
    * fix overflow avx2::v_pack(uint32)
    * implement two universal intrinsics v_expand_low and v_expand_high
This commit is contained in:
Sayed Adel
2018-09-12 21:32:34 +02:00
parent 1ff11c84ab
commit 5771fd693d
13 changed files with 744 additions and 193 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/video/src/lkpyramid.cpp
+2 -2
View File
@@ -93,7 +93,7 @@ static void calcSharrDeriv(const cv::Mat& src, cv::Mat& dst)
v_int16x8 s2 = v_reinterpret_as_s16(v_load_expand(srow2 + x));
v_int16x8 t1 = s2 - s0;
v_int16x8 t0 = (s0 + s2) * c3 + s1 * c10;
v_int16x8 t0 = v_mul_wrap(s0 + s2, c3) + v_mul_wrap(s1, c10);
v_store(trow0 + x, t0);
v_store(trow1 + x, t1);
@@ -131,7 +131,7 @@ static void calcSharrDeriv(const cv::Mat& src, cv::Mat& dst)
v_int16x8 s4 = v_load(trow1 + x + cn);
v_int16x8 t0 = s1 - s0;
v_int16x8 t1 = ((s2 + s4) * c3) + (s3 * c10);
v_int16x8 t1 = v_mul_wrap(s2 + s4, c3) + v_mul_wrap(s3, c10);
v_store_interleave((drow + x*2), t0, t1);
}