Skip to content

fix: make simd4f::sum() consistent across all SIMD paths #3172

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
davisking merged 1 commit into from
Oct 17, 2026
Merged

fix: make simd4f::sum() consistent across all SIMD paths #3172

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 5 additions & 4 deletions dlib/simd/simd4f.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -559,14 +559,15 @@ namespace dlib
simd4f temp = _mm_hadd_ps(item,item);
return _mm_cvtss_f32(_mm_hadd_ps(temp,temp));
#elif defined(DLIB_HAVE_SSE2) && (!defined(_MSC_VER) || _MSC_VER!=1400)
simd4f temp = _mm_add_ps(item,_mm_movehl_ps(item,item));
simd4f temp2 = _mm_shuffle_ps(temp,temp,1);
simd4f temp = _mm_shuffle_ps(item, item, _MM_SHUFFLE(2, 3, 0, 1));
temp = _mm_add_ps(temp, item);
simd4f temp2 = _mm_shuffle_ps(temp,temp,_MM_SHUFFLE(0, 1, 2, 3));
return _mm_cvtss_f32(_mm_add_ss(temp,temp2));
#elif defined(DLIB_HAVE_NEON)
float32x2_t r = vadd_f32(vget_high_f32(item), vget_low_f32(item));
float32x2_t r = vpadd_f32(vget_high_f32(item), vget_low_f32(item));
return vget_lane_f32(vpadd_f32(r, r), 0);
#else
return item[0]+item[1]+item[2]+item[3];
return (item[0]+item[1])+(item[2]+item[3]);
#endif
}

Expand Down
Loading