core: improve norm of hal rvv by fengyuentau · Pull Request #26991 · opencv/opencv

fengyuentau · 2025-02-28T10:36:09Z

Merge with opencv/opencv_extra#1241

Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

I agree to contribute to the project under Apache 2 License.
To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
The PR is proposed to the proper branch
There is a reference to the original bug report and related work
There is accuracy test, performance test and test data in opencv_extra repository, if applicable
Patch to opencv_extra has the same branch name.
The feature is well documented and sample code can be built with the project CMake

modules/core/src/hal_replacement.hpp

fengyuentau · 2025-03-08T16:21:30Z

Conflicts are resolved. Feel free to review @asmorkalov

fengyuentau · 2025-03-09T17:12:08Z

Updated perf results: perf-norm.zip

asmorkalov · 2025-03-10T16:32:05Z

modules/core/src/hal_replacement.hpp

-inline int hal_ni_norm(const uchar* src, size_t src_step, const uchar* mask, size_t mask_step, int width,
-                       int height, int type, int norm_type, double* result) { return CV_HAL_ERROR_NOT_IMPLEMENTED; }


FastCV HAL implements the old API too.

opencv-alalek · 2025-03-14T06:58:50Z

modules/core/src/norm.simd.hpp

 #include "precomp.hpp"

-#if CV_RVV
-#include "norm.rvv1p0.hpp"


Where is replacement for these removed optimizations? (added by #26885)

All moved to hal rvv.

…implementation with rvv covering different data types, channels and norm types

fengyuentau · 2025-03-18T06:05:21Z

I refactored this PR with the same handling of overflow accumulator.

asmorkalov · 2025-03-18T06:29:08Z

The patch birngs speedup for many cases, but there are some strange regressions against current 4.x:

Details

norm2::Size_MatType_NormType::(127x61, 8UC1, NORM_INF) 	0.004 	0.003 	1.11
norm2::Size_MatType_NormType::(127x61, 8UC1, NORM_INF|NORM_RELATIVE) 	0.004 	0.004 	1.21
norm2::Size_MatType_NormType::(127x61, 8UC1, NORM_L1) 	0.004 	0.004 	1.02
norm2::Size_MatType_NormType::(127x61, 8UC1, NORM_L1|NORM_RELATIVE) 	0.006 	0.005 	1.15
norm2::Size_MatType_NormType::(127x61, 8UC1, NORM_L2) 	0.005 	0.005 	1.01
norm2::Size_MatType_NormType::(127x61, 8UC1, NORM_L2|NORM_RELATIVE) 	0.009 	0.007 	1.31
norm2::Size_MatType_NormType::(127x61, 8SC1, NORM_INF) 	0.004 	0.003 	1.05
norm2::Size_MatType_NormType::(127x61, 8SC1, NORM_INF|NORM_RELATIVE) 	0.007 	0.005 	1.28
norm2::Size_MatType_NormType::(127x61, 8SC1, NORM_L1) 	0.006 	0.005 	1.07
norm2::Size_MatType_NormType::(127x61, 8SC1, NORM_L1|NORM_RELATIVE) 	0.009 	0.008 	1.16
norm2::Size_MatType_NormType::(127x61, 8SC1, NORM_L2) 	0.006 	0.005 	1.06
norm2::Size_MatType_NormType::(127x61, 8SC1, NORM_L2|NORM_RELATIVE) 	0.008 	0.008 	1.05
norm2::Size_MatType_NormType::(127x61, 16UC1, NORM_INF) 	0.007 	0.006 	1.14
norm2::Size_MatType_NormType::(127x61, 16UC1, NORM_INF|NORM_RELATIVE) 	0.010 	0.009 	1.07
norm2::Size_MatType_NormType::(127x61, 16UC1, NORM_L1) 	0.008 	0.007 	1.07
norm2::Size_MatType_NormType::(127x61, 16UC1, NORM_L1|NORM_RELATIVE) 	0.011 	0.010 	1.09
norm2::Size_MatType_NormType::(127x61, 16UC1, NORM_L2) 	0.012 	0.037 	0.32
norm2::Size_MatType_NormType::(127x61, 16UC1, NORM_L2|NORM_RELATIVE) 	0.019 	0.043 	0.45
norm2::Size_MatType_NormType::(127x61, 16SC1, NORM_INF) 	0.009 	0.008 	1.08
norm2::Size_MatType_NormType::(127x61, 16SC1, NORM_INF|NORM_RELATIVE) 	0.013 	0.012 	1.16
norm2::Size_MatType_NormType::(127x61, 16SC1, NORM_L1) 	0.009 	0.009 	1.06
norm2::Size_MatType_NormType::(127x61, 16SC1, NORM_L1|NORM_RELATIVE) 	0.013 	0.013 	1.01
norm2::Size_MatType_NormType::(127x61, 16SC1, NORM_L2) 	0.013 	0.037 	0.35
norm2::Size_MatType_NormType::(127x61, 16SC1, NORM_L2|NORM_RELATIVE) 	0.021 	0.043 	0.48
norm2::Size_MatType_NormType::(127x61, 32SC1, NORM_INF) 	0.014 	0.014 	1.00
norm2::Size_MatType_NormType::(127x61, 32SC1, NORM_INF|NORM_RELATIVE) 	0.022 	0.020 	1.15
norm2::Size_MatType_NormType::(127x61, 32SC1, NORM_L1) 	0.012 	0.035 	0.35
norm2::Size_MatType_NormType::(127x61, 32SC1, NORM_L1|NORM_RELATIVE) 	0.022 	0.044 	0.49
norm2::Size_MatType_NormType::(127x61, 32SC1, NORM_L2) 	0.012 	0.036 	0.34
norm2::Size_MatType_NormType::(127x61, 32SC1, NORM_L2|NORM_RELATIVE) 	0.029 	0.044 	0.66
norm2::Size_MatType_NormType::(127x61, 32FC1, NORM_INF) 	0.010 	0.009 	1.04
norm2::Size_MatType_NormType::(127x61, 32FC1, NORM_INF|NORM_RELATIVE) 	0.013 	0.013 	1.03
norm2::Size_MatType_NormType::(127x61, 32FC1, NORM_L1) 	0.011 	0.010 	1.06
norm2::Size_MatType_NormType::(127x61, 32FC1, NORM_L1|NORM_RELATIVE) 	0.017 	0.016 	1.09
norm2::Size_MatType_NormType::(127x61, 32FC1, NORM_L2) 	0.010 	0.010 	1.02
norm2::Size_MatType_NormType::(127x61, 32FC1, NORM_L2|NORM_RELATIVE) 	0.016 	0.015 	1.08
norm2::Size_MatType_NormType::(127x61, 64FC1, NORM_INF) 	0.018 	0.091 	0.20
norm2::Size_MatType_NormType::(127x61, 64FC1, NORM_INF|NORM_RELATIVE) 	0.030 	0.099 	0.30
norm2::Size_MatType_NormType::(127x61, 64FC1, NORM_L1) 	0.018 	0.032 	0.55
norm2::Size_MatType_NormType::(127x61, 64FC1, NORM_L1|NORM_RELATIVE) 	0.028 	0.041 	0.69
norm2::Size_MatType_NormType::(127x61, 64FC1, NORM_L2) 	0.019 	0.033 	0.57
norm2::Size_MatType_NormType::(127x61, 64FC1, NORM_L2|NORM_RELATIVE) 	0.029 	0.043 	0.69
norm2::Size_MatType_NormType::(127x61, 8UC4, NORM_INF) 	0.010 	0.010 	1.00
norm2::Size_MatType_NormType::(127x61, 8UC4, NORM_INF|NORM_RELATIVE) 	0.013 	0.013 	1.01
norm2::Size_MatType_NormType::(127x61, 8UC4, NORM_L1) 	0.015 	0.015 	0.99
norm2::Size_MatType_NormType::(127x61, 8UC4, NORM_L1|NORM_RELATIVE) 	0.023 	0.019 	1.19
norm2::Size_MatType_NormType::(127x61, 8UC4, NORM_L2) 	0.017 	0.016 	1.02
norm2::Size_MatType_NormType::(127x61, 8UC4, NORM_L2|NORM_RELATIVE) 	0.028 	0.023 	1.21
norm2::Size_MatType_NormType::(640x480, 8UC1, NORM_INF) 	0.125 	0.125 	1.00
norm2::Size_MatType_NormType::(640x480, 8UC1, NORM_INF|NORM_RELATIVE) 	0.168 	0.168 	1.00
norm2::Size_MatType_NormType::(640x480, 8UC1, NORM_L1) 	0.129 	0.132 	0.98
norm2::Size_MatType_NormType::(640x480, 8UC1, NORM_L1|NORM_RELATIVE) 	0.210 	0.180 	1.16
norm2::Size_MatType_NormType::(640x480, 8UC1, NORM_L2) 	0.152 	0.153 	1.00
norm2::Size_MatType_NormType::(640x480, 8UC1, NORM_L2|NORM_RELATIVE) 	0.258 	0.215 	1.20
norm2::Size_MatType_NormType::(640x480, 8SC1, NORM_INF) 	0.122 	0.121 	1.00
norm2::Size_MatType_NormType::(640x480, 8SC1, NORM_INF|NORM_RELATIVE) 	0.205 	0.170 	1.21
norm2::Size_MatType_NormType::(640x480, 8SC1, NORM_L1) 	0.183 	0.179 	1.02
norm2::Size_MatType_NormType::(640x480, 8SC1, NORM_L1|NORM_RELATIVE) 	0.284 	0.254 	1.12
norm2::Size_MatType_NormType::(640x480, 8SC1, NORM_L2) 	0.182 	0.178 	1.03
norm2::Size_MatType_NormType::(640x480, 8SC1, NORM_L2|NORM_RELATIVE) 	0.248 	0.242 	1.03
norm2::Size_MatType_NormType::(640x480, 16UC1, NORM_INF) 	0.186 	0.183 	1.02
norm2::Size_MatType_NormType::(640x480, 16UC1, NORM_INF|NORM_RELATIVE) 	0.274 	0.273 	1.01
norm2::Size_MatType_NormType::(640x480, 16UC1, NORM_L1) 	0.215 	0.213 	1.01
norm2::Size_MatType_NormType::(640x480, 16UC1, NORM_L1|NORM_RELATIVE) 	0.309 	0.299 	1.03
norm2::Size_MatType_NormType::(640x480, 16UC1, NORM_L2) 	0.316 	1.377 	0.23
norm2::Size_MatType_NormType::(640x480, 16UC1, NORM_L2|NORM_RELATIVE) 	0.552 	1.594 	0.35
norm2::Size_MatType_NormType::(640x480, 16SC1, NORM_INF) 	0.231 	0.232 	1.00
norm2::Size_MatType_NormType::(640x480, 16SC1, NORM_INF|NORM_RELATIVE) 	0.396 	0.329 	1.20
norm2::Size_MatType_NormType::(640x480, 16SC1, NORM_L1) 	0.274 	0.270 	1.01
norm2::Size_MatType_NormType::(640x480, 16SC1, NORM_L1|NORM_RELATIVE) 	0.403 	0.386 	1.04
norm2::Size_MatType_NormType::(640x480, 16SC1, NORM_L2) 	0.370 	1.377 	0.27
norm2::Size_MatType_NormType::(640x480, 16SC1, NORM_L2|NORM_RELATIVE) 	0.607 	1.587 	0.38
norm2::Size_MatType_NormType::(640x480, 32SC1, NORM_INF) 	0.542 	0.542 	1.00
norm2::Size_MatType_NormType::(640x480, 32SC1, NORM_INF|NORM_RELATIVE) 	0.794 	0.739 	1.08
norm2::Size_MatType_NormType::(640x480, 32SC1, NORM_L1) 	0.550 	1.333 	0.41
norm2::Size_MatType_NormType::(640x480, 32SC1, NORM_L1|NORM_RELATIVE) 	0.879 	1.615 	0.54
norm2::Size_MatType_NormType::(640x480, 32SC1, NORM_L2) 	0.548 	1.377 	0.40
norm2::Size_MatType_NormType::(640x480, 32SC1, NORM_L2|NORM_RELATIVE) 	1.137 	1.630 	0.70
norm2::Size_MatType_NormType::(640x480, 32FC1, NORM_INF) 	0.573 	0.574 	1.00
norm2::Size_MatType_NormType::(640x480, 32FC1, NORM_INF|NORM_RELATIVE) 	0.734 	0.733 	1.00
norm2::Size_MatType_NormType::(640x480, 32FC1, NORM_L1) 	0.556 	0.539 	1.03
norm2::Size_MatType_NormType::(640x480, 32FC1, NORM_L1|NORM_RELATIVE) 	0.778 	0.755 	1.03
norm2::Size_MatType_NormType::(640x480, 32FC1, NORM_L2) 	0.549 	0.542 	1.01
norm2::Size_MatType_NormType::(640x480, 32FC1, NORM_L2|NORM_RELATIVE) 	0.768 	0.723 	1.06
norm2::Size_MatType_NormType::(640x480, 64FC1, NORM_INF) 	0.790 	3.581 	0.22
norm2::Size_MatType_NormType::(640x480, 64FC1, NORM_INF|NORM_RELATIVE) 	1.193 	3.936 	0.30
norm2::Size_MatType_NormType::(640x480, 64FC1, NORM_L1) 	0.790 	1.229 	0.64
norm2::Size_MatType_NormType::(640x480, 64FC1, NORM_L1|NORM_RELATIVE) 	1.163 	1.570 	0.74
norm2::Size_MatType_NormType::(640x480, 64FC1, NORM_L2) 	0.792 	1.281 	0.62
norm2::Size_MatType_NormType::(640x480, 64FC1, NORM_L2|NORM_RELATIVE) 	1.151 	1.614 	0.71
norm2::Size_MatType_NormType::(640x480, 8UC4, NORM_INF) 	0.567 	0.574 	0.99
norm2::Size_MatType_NormType::(640x480, 8UC4, NORM_INF|NORM_RELATIVE) 	0.734 	0.733 	1.00
norm2::Size_MatType_NormType::(640x480, 8UC4, NORM_L1) 	0.569 	0.575 	0.99
norm2::Size_MatType_NormType::(640x480, 8UC4, NORM_L1|NORM_RELATIVE) 	0.898 	0.752 	1.19
norm2::Size_MatType_NormType::(640x480, 8UC4, NORM_L2) 	0.591 	0.596 	0.99
norm2::Size_MatType_NormType::(640x480, 8UC4, NORM_L2|NORM_RELATIVE) 	1.010 	0.827 	1.22
norm2::Size_MatType_NormType::(1280x720, 8UC1, NORM_INF) 	0.359 	0.357 	1.01
norm2::Size_MatType_NormType::(1280x720, 8UC1, NORM_INF|NORM_RELATIVE) 	0.467 	0.472 	0.99
norm2::Size_MatType_NormType::(1280x720, 8UC1, NORM_L1) 	0.369 	0.381 	0.97
norm2::Size_MatType_NormType::(1280x720, 8UC1, NORM_L1|NORM_RELATIVE) 	0.610 	0.515 	1.18
norm2::Size_MatType_NormType::(1280x720, 8UC1, NORM_L2) 	0.434 	0.439 	0.99
norm2::Size_MatType_NormType::(1280x720, 8UC1, NORM_L2|NORM_RELATIVE) 	0.754 	0.613 	1.23
norm2::Size_MatType_NormType::(1280x720, 8SC1, NORM_INF) 	0.349 	0.344 	1.02
norm2::Size_MatType_NormType::(1280x720, 8SC1, NORM_INF|NORM_RELATIVE) 	0.583 	0.485 	1.20
norm2::Size_MatType_NormType::(1280x720, 8SC1, NORM_L1) 	0.523 	0.514 	1.02
norm2::Size_MatType_NormType::(1280x720, 8SC1, NORM_L1|NORM_RELATIVE) 	0.831 	0.728 	1.14
norm2::Size_MatType_NormType::(1280x720, 8SC1, NORM_L2) 	0.525 	0.509 	1.03
norm2::Size_MatType_NormType::(1280x720, 8SC1, NORM_L2|NORM_RELATIVE) 	0.714 	0.695 	1.03
norm2::Size_MatType_NormType::(1280x720, 16UC1, NORM_INF) 	0.563 	0.575 	0.98
norm2::Size_MatType_NormType::(1280x720, 16UC1, NORM_INF|NORM_RELATIVE) 	0.822 	0.836 	0.98
norm2::Size_MatType_NormType::(1280x720, 16UC1, NORM_L1) 	0.657 	0.647 	1.01
norm2::Size_MatType_NormType::(1280x720, 16UC1, NORM_L1|NORM_RELATIVE) 	0.931 	0.936 	0.99
norm2::Size_MatType_NormType::(1280x720, 16UC1, NORM_L2) 	0.928 	4.125 	0.22
norm2::Size_MatType_NormType::(1280x720, 16UC1, NORM_L2|NORM_RELATIVE) 	1.633 	4.736 	0.34
norm2::Size_MatType_NormType::(1280x720, 16SC1, NORM_INF) 	0.683 	0.690 	0.99
norm2::Size_MatType_NormType::(1280x720, 16SC1, NORM_INF|NORM_RELATIVE) 	1.165 	0.971 	1.20
norm2::Size_MatType_NormType::(1280x720, 16SC1, NORM_L1) 	0.805 	0.806 	1.00
norm2::Size_MatType_NormType::(1280x720, 16SC1, NORM_L1|NORM_RELATIVE) 	1.185 	1.136 	1.04
norm2::Size_MatType_NormType::(1280x720, 16SC1, NORM_L2) 	1.092 	4.122 	0.27
norm2::Size_MatType_NormType::(1280x720, 16SC1, NORM_L2|NORM_RELATIVE) 	1.791 	4.719 	0.38
norm2::Size_MatType_NormType::(1280x720, 32SC1, NORM_INF) 	1.486 	1.374 	1.08
norm2::Size_MatType_NormType::(1280x720, 32SC1, NORM_INF|NORM_RELATIVE) 	2.300 	2.020 	1.14
norm2::Size_MatType_NormType::(1280x720, 32SC1, NORM_L1) 	1.464 	3.971 	0.37
norm2::Size_MatType_NormType::(1280x720, 32SC1, NORM_L1|NORM_RELATIVE) 	2.350 	4.827 	0.49
norm2::Size_MatType_NormType::(1280x720, 32SC1, NORM_L2) 	1.482 	4.131 	0.36
norm2::Size_MatType_NormType::(1280x720, 32SC1, NORM_L2|NORM_RELATIVE) 	3.169 	4.820 	0.66
norm2::Size_MatType_NormType::(1280x720, 32FC1, NORM_INF) 	1.521 	1.526 	1.00
norm2::Size_MatType_NormType::(1280x720, 32FC1, NORM_INF|NORM_RELATIVE) 	2.014 	2.031 	0.99
norm2::Size_MatType_NormType::(1280x720, 32FC1, NORM_L1) 	1.619 	1.287 	1.26
norm2::Size_MatType_NormType::(1280x720, 32FC1, NORM_L1|NORM_RELATIVE) 	2.275 	1.937 	1.17
norm2::Size_MatType_NormType::(1280x720, 32FC1, NORM_L2) 	1.646 	1.296 	1.27
norm2::Size_MatType_NormType::(1280x720, 32FC1, NORM_L2|NORM_RELATIVE) 	2.287 	1.881 	1.22
norm2::Size_MatType_NormType::(1280x720, 64FC1, NORM_INF) 	2.436 	10.621 	0.23
norm2::Size_MatType_NormType::(1280x720, 64FC1, NORM_INF|NORM_RELATIVE) 	3.652 	11.623 	0.31
norm2::Size_MatType_NormType::(1280x720, 64FC1, NORM_L1) 	2.418 	3.659 	0.66
norm2::Size_MatType_NormType::(1280x720, 64FC1, NORM_L1|NORM_RELATIVE) 	3.530 	4.680 	0.75
norm2::Size_MatType_NormType::(1280x720, 64FC1, NORM_L2) 	2.389 	3.815 	0.63
norm2::Size_MatType_NormType::(1280x720, 64FC1, NORM_L2|NORM_RELATIVE) 	3.457 	4.845 	0.71
norm2::Size_MatType_NormType::(1280x720, 8UC4, NORM_INF) 	1.495 	1.503 	0.99
norm2::Size_MatType_NormType::(1280x720, 8UC4, NORM_INF|NORM_RELATIVE) 	2.010 	2.003 	1.00
norm2::Size_MatType_NormType::(1280x720, 8UC4, NORM_L1) 	1.545 	1.577 	0.98
norm2::Size_MatType_NormType::(1280x720, 8UC4, NORM_L1|NORM_RELATIVE) 	2.509 	2.217 	1.13
norm2::Size_MatType_NormType::(1280x720, 8UC4, NORM_L2) 	1.739 	1.767 	0.98
norm2::Size_MatType_NormType::(1280x720, 8UC4, NORM_L2|NORM_RELATIVE) 	3.010 	2.503 	1.20
norm2::Size_MatType_NormType::(1920x1080, 8UC1, NORM_INF) 	0.602 	0.621 	0.97
norm2::Size_MatType_NormType::(1920x1080, 8UC1, NORM_INF|NORM_RELATIVE) 	0.872 	0.894 	0.98
norm2::Size_MatType_NormType::(1920x1080, 8UC1, NORM_L1) 	0.866 	0.926 	0.93
norm2::Size_MatType_NormType::(1920x1080, 8UC1, NORM_L1|NORM_RELATIVE) 	1.411 	1.233 	1.14
norm2::Size_MatType_NormType::(1920x1080, 8UC1, NORM_L2) 	0.970 	0.982 	0.99
norm2::Size_MatType_NormType::(1920x1080, 8UC1, NORM_L2|NORM_RELATIVE) 	1.684 	1.388 	1.21
norm2::Size_MatType_NormType::(1920x1080, 8SC1, NORM_INF) 	0.773 	0.788 	0.98
norm2::Size_MatType_NormType::(1920x1080, 8SC1, NORM_INF|NORM_RELATIVE) 	1.318 	1.112 	1.18
norm2::Size_MatType_NormType::(1920x1080, 8SC1, NORM_L1) 	1.180 	1.149 	1.03
norm2::Size_MatType_NormType::(1920x1080, 8SC1, NORM_L1|NORM_RELATIVE) 	1.858 	1.621 	1.15
norm2::Size_MatType_NormType::(1920x1080, 8SC1, NORM_L2) 	1.177 	1.148 	1.03
norm2::Size_MatType_NormType::(1920x1080, 8SC1, NORM_L2|NORM_RELATIVE) 	1.604 	1.563 	1.03
norm2::Size_MatType_NormType::(1920x1080, 16UC1, NORM_INF) 	1.313 	1.260 	1.04
norm2::Size_MatType_NormType::(1920x1080, 16UC1, NORM_INF|NORM_RELATIVE) 	1.893 	1.842 	1.03
norm2::Size_MatType_NormType::(1920x1080, 16UC1, NORM_L1) 	1.519 	1.451 	1.05
norm2::Size_MatType_NormType::(1920x1080, 16UC1, NORM_L1|NORM_RELATIVE) 	2.168 	2.035 	1.07
norm2::Size_MatType_NormType::(1920x1080, 16UC1, NORM_L2) 	2.195 	9.267 	0.24
norm2::Size_MatType_NormType::(1920x1080, 16UC1, NORM_L2|NORM_RELATIVE) 	3.729 	10.621 	0.35
norm2::Size_MatType_NormType::(1920x1080, 16SC1, NORM_INF) 	1.627 	1.567 	1.04
norm2::Size_MatType_NormType::(1920x1080, 16SC1, NORM_INF|NORM_RELATIVE) 	2.713 	2.172 	1.25
norm2::Size_MatType_NormType::(1920x1080, 16SC1, NORM_L1) 	1.907 	1.846 	1.03
norm2::Size_MatType_NormType::(1920x1080, 16SC1, NORM_L1|NORM_RELATIVE) 	2.822 	2.591 	1.09
norm2::Size_MatType_NormType::(1920x1080, 16SC1, NORM_L2) 	2.545 	9.266 	0.27
norm2::Size_MatType_NormType::(1920x1080, 16SC1, NORM_L2|NORM_RELATIVE) 	4.095 	10.607 	0.39
norm2::Size_MatType_NormType::(1920x1080, 32SC1, NORM_INF) 	3.255 	3.064 	1.06
norm2::Size_MatType_NormType::(1920x1080, 32SC1, NORM_INF|NORM_RELATIVE) 	5.028 	4.524 	1.11
norm2::Size_MatType_NormType::(1920x1080, 32SC1, NORM_L1) 	3.365 	8.909 	0.38
norm2::Size_MatType_NormType::(1920x1080, 32SC1, NORM_L1|NORM_RELATIVE) 	5.393 	10.817 	0.50
norm2::Size_MatType_NormType::(1920x1080, 32SC1, NORM_L2) 	3.392 	9.241 	0.37
norm2::Size_MatType_NormType::(1920x1080, 32SC1, NORM_L2|NORM_RELATIVE) 	7.168 	10.806 	0.66
norm2::Size_MatType_NormType::(1920x1080, 32FC1, NORM_INF) 	3.453 	3.516 	0.98
norm2::Size_MatType_NormType::(1920x1080, 32FC1, NORM_INF|NORM_RELATIVE) 	4.587 	4.629 	0.99
norm2::Size_MatType_NormType::(1920x1080, 32FC1, NORM_L1) 	3.894 	3.268 	1.19
norm2::Size_MatType_NormType::(1920x1080, 32FC1, NORM_L1|NORM_RELATIVE) 	5.361 	4.725 	1.13
norm2::Size_MatType_NormType::(1920x1080, 32FC1, NORM_L2) 	3.923 	3.392 	1.16
norm2::Size_MatType_NormType::(1920x1080, 32FC1, NORM_L2|NORM_RELATIVE) 	5.326 	4.612 	1.15
norm2::Size_MatType_NormType::(1920x1080, 64FC1, NORM_INF) 	5.179 	23.924 	0.22
norm2::Size_MatType_NormType::(1920x1080, 64FC1, NORM_INF|NORM_RELATIVE) 	7.949 	26.060 	0.31
norm2::Size_MatType_NormType::(1920x1080, 64FC1, NORM_L1) 	5.201 	8.201 	0.63
norm2::Size_MatType_NormType::(1920x1080, 64FC1, NORM_L1|NORM_RELATIVE) 	7.606 	10.436 	0.73
norm2::Size_MatType_NormType::(1920x1080, 64FC1, NORM_L2) 	5.181 	8.539 	0.61
norm2::Size_MatType_NormType::(1920x1080, 64FC1, NORM_L2|NORM_RELATIVE) 	7.599 	10.700 	0.71
norm2::Size_MatType_NormType::(1920x1080, 8UC4, NORM_INF) 	3.454 	3.491 	0.99
norm2::Size_MatType_NormType::(1920x1080, 8UC4, NORM_INF|NORM_RELATIVE) 	4.533 	4.569 	0.99
norm2::Size_MatType_NormType::(1920x1080, 8UC4, NORM_L1) 	3.787 	3.636 	1.04
norm2::Size_MatType_NormType::(1920x1080, 8UC4, NORM_L1|NORM_RELATIVE) 	5.941 	4.851 	1.22
norm2::Size_MatType_NormType::(1920x1080, 8UC4, NORM_L2) 	3.873 	3.948 	0.98
norm2::Size_MatType_NormType::(1920x1080, 8UC4, NORM_L2|NORM_RELATIVE) 	6.692 	5.516 	1.21
norm2_mask::Size_MatType_NormType::(127x61, 8UC1, NORM_INF) 	0.008 	0.008 	1.01
norm2_mask::Size_MatType_NormType::(127x61, 8UC1, NORM_INF|NORM_RELATIVE) 	0.010 	0.009 	1.15
norm2_mask::Size_MatType_NormType::(127x61, 8UC1, NORM_L1) 	0.006 	0.006 	1.01
norm2_mask::Size_MatType_NormType::(127x61, 8UC1, NORM_L1|NORM_RELATIVE) 	0.010 	0.008 	1.23
norm2_mask::Size_MatType_NormType::(127x61, 8UC1, NORM_L2) 	0.008 	0.008 	1.01
norm2_mask::Size_MatType_NormType::(127x61, 8UC1, NORM_L2|NORM_RELATIVE) 	0.013 	0.011 	1.23
norm2_mask::Size_MatType_NormType::(127x61, 8SC1, NORM_INF) 	0.181 	0.181 	1.00
norm2_mask::Size_MatType_NormType::(127x61, 8SC1, NORM_INF|NORM_RELATIVE) 	0.337 	0.189 	1.78
norm2_mask::Size_MatType_NormType::(127x61, 8SC1, NORM_L1) 	0.200 	0.200 	1.00
norm2_mask::Size_MatType_NormType::(127x61, 8SC1, NORM_L1|NORM_RELATIVE) 	0.378 	0.209 	1.81
norm2_mask::Size_MatType_NormType::(127x61, 8SC1, NORM_L2) 	0.190 	0.191 	1.00
norm2_mask::Size_MatType_NormType::(127x61, 8SC1, NORM_L2|NORM_RELATIVE) 	0.343 	0.199 	1.72
norm2_mask::Size_MatType_NormType::(127x61, 16UC1, NORM_INF) 	0.178 	0.178 	1.00
norm2_mask::Size_MatType_NormType::(127x61, 16UC1, NORM_INF|NORM_RELATIVE) 	0.306 	0.186 	1.64
norm2_mask::Size_MatType_NormType::(127x61, 16UC1, NORM_L1) 	0.205 	0.205 	1.00
norm2_mask::Size_MatType_NormType::(127x61, 16UC1, NORM_L1|NORM_RELATIVE) 	0.352 	0.214 	1.64
norm2_mask::Size_MatType_NormType::(127x61, 16UC1, NORM_L2) 	0.173 	0.146 	1.19
norm2_mask::Size_MatType_NormType::(127x61, 16UC1, NORM_L2|NORM_RELATIVE) 	0.330 	0.160 	2.06
norm2_mask::Size_MatType_NormType::(127x61, 16SC1, NORM_INF) 	0.177 	0.177 	1.00
norm2_mask::Size_MatType_NormType::(127x61, 16SC1, NORM_INF|NORM_RELATIVE) 	0.335 	0.188 	1.78
norm2_mask::Size_MatType_NormType::(127x61, 16SC1, NORM_L1) 	0.205 	0.205 	1.00
norm2_mask::Size_MatType_NormType::(127x61, 16SC1, NORM_L1|NORM_RELATIVE) 	0.376 	0.215 	1.75
norm2_mask::Size_MatType_NormType::(127x61, 16SC1, NORM_L2) 	0.173 	0.145 	1.19
norm2_mask::Size_MatType_NormType::(127x61, 16SC1, NORM_L2|NORM_RELATIVE) 	0.236 	0.160 	1.47
norm2_mask::Size_MatType_NormType::(127x61, 32SC1, NORM_INF) 	0.182 	0.182 	1.00
norm2_mask::Size_MatType_NormType::(127x61, 32SC1, NORM_INF|NORM_RELATIVE) 	0.329 	0.196 	1.68
norm2_mask::Size_MatType_NormType::(127x61, 32SC1, NORM_L1) 	0.176 	0.155 	1.14
norm2_mask::Size_MatType_NormType::(127x61, 32SC1, NORM_L1|NORM_RELATIVE) 	0.325 	0.172 	1.88
norm2_mask::Size_MatType_NormType::(127x61, 32SC1, NORM_L2) 	0.173 	0.151 	1.15
norm2_mask::Size_MatType_NormType::(127x61, 32SC1, NORM_L2|NORM_RELATIVE) 	0.315 	0.168 	1.87
norm2_mask::Size_MatType_NormType::(127x61, 32FC1, NORM_INF) 	0.012 	0.011 	1.05
norm2_mask::Size_MatType_NormType::(127x61, 32FC1, NORM_INF|NORM_RELATIVE) 	0.017 	0.016 	1.06
norm2_mask::Size_MatType_NormType::(127x61, 32FC1, NORM_L1) 	0.014 	0.013 	1.04
norm2_mask::Size_MatType_NormType::(127x61, 32FC1, NORM_L1|NORM_RELATIVE) 	0.022 	0.021 	1.03
norm2_mask::Size_MatType_NormType::(127x61, 32FC1, NORM_L2) 	0.013 	0.012 	1.04
norm2_mask::Size_MatType_NormType::(127x61, 32FC1, NORM_L2|NORM_RELATIVE) 	0.020 	0.019 	1.04
norm2_mask::Size_MatType_NormType::(127x61, 64FC1, NORM_INF) 	0.124 	0.118 	1.05
norm2_mask::Size_MatType_NormType::(127x61, 64FC1, NORM_INF|NORM_RELATIVE) 	0.220 	0.133 	1.65
norm2_mask::Size_MatType_NormType::(127x61, 64FC1, NORM_L1) 	0.153 	0.134 	1.14
norm2_mask::Size_MatType_NormType::(127x61, 64FC1, NORM_L1|NORM_RELATIVE) 	0.217 	0.149 	1.46
norm2_mask::Size_MatType_NormType::(127x61, 64FC1, NORM_L2) 	0.159 	0.139 	1.14
norm2_mask::Size_MatType_NormType::(127x61, 64FC1, NORM_L2|NORM_RELATIVE) 	0.225 	0.156 	1.44
norm2_mask::Size_MatType_NormType::(127x61, 8UC4, NORM_INF) 	0.017 	0.017 	1.03
norm2_mask::Size_MatType_NormType::(127x61, 8UC4, NORM_INF|NORM_RELATIVE) 	0.023 	0.022 	1.03
norm2_mask::Size_MatType_NormType::(127x61, 8UC4, NORM_L1) 	0.023 	0.021 	1.08
norm2_mask::Size_MatType_NormType::(127x61, 8UC4, NORM_L1|NORM_RELATIVE) 	0.039 	0.029 	1.35
norm2_mask::Size_MatType_NormType::(127x61, 8UC4, NORM_L2) 	0.027 	0.026 	1.06
norm2_mask::Size_MatType_NormType::(127x61, 8UC4, NORM_L2|NORM_RELATIVE) 	0.048 	0.038 	1.25
norm2_mask::Size_MatType_NormType::(640x480, 8UC1, NORM_INF) 	0.184 	0.183 	1.01
norm2_mask::Size_MatType_NormType::(640x480, 8UC1, NORM_INF|NORM_RELATIVE) 	0.268 	0.266 	1.00
norm2_mask::Size_MatType_NormType::(640x480, 8UC1, NORM_L1) 	0.202 	0.198 	1.02
norm2_mask::Size_MatType_NormType::(640x480, 8UC1, NORM_L1|NORM_RELATIVE) 	0.336 	0.287 	1.17
norm2_mask::Size_MatType_NormType::(640x480, 8UC1, NORM_L2) 	0.228 	0.226 	1.01
norm2_mask::Size_MatType_NormType::(640x480, 8UC1, NORM_L2|NORM_RELATIVE) 	0.394 	0.335 	1.17
norm2_mask::Size_MatType_NormType::(640x480, 8SC1, NORM_INF) 	7.176 	7.200 	1.00
norm2_mask::Size_MatType_NormType::(640x480, 8SC1, NORM_INF|NORM_RELATIVE) 	13.255 	7.503 	1.77
norm2_mask::Size_MatType_NormType::(640x480, 8SC1, NORM_L1) 	7.907 	7.895 	1.00
norm2_mask::Size_MatType_NormType::(640x480, 8SC1, NORM_L1|NORM_RELATIVE) 	14.895 	8.249 	1.81
norm2_mask::Size_MatType_NormType::(640x480, 8SC1, NORM_L2) 	7.513 	7.526 	1.00
norm2_mask::Size_MatType_NormType::(640x480, 8SC1, NORM_L2|NORM_RELATIVE) 	13.568 	7.879 	1.72
norm2_mask::Size_MatType_NormType::(640x480, 16UC1, NORM_INF) 	7.012 	6.998 	1.00
norm2_mask::Size_MatType_NormType::(640x480, 16UC1, NORM_INF|NORM_RELATIVE) 	12.042 	7.338 	1.64
norm2_mask::Size_MatType_NormType::(640x480, 16UC1, NORM_L1) 	8.105 	8.112 	1.00
norm2_mask::Size_MatType_NormType::(640x480, 16UC1, NORM_L1|NORM_RELATIVE) 	13.878 	8.472 	1.64
norm2_mask::Size_MatType_NormType::(640x480, 16UC1, NORM_L2) 	6.825 	5.732 	1.19
norm2_mask::Size_MatType_NormType::(640x480, 16UC1, NORM_L2|NORM_RELATIVE) 	13.041 	6.263 	2.08
norm2_mask::Size_MatType_NormType::(640x480, 16SC1, NORM_INF) 	7.005 	7.007 	1.00
norm2_mask::Size_MatType_NormType::(640x480, 16SC1, NORM_INF|NORM_RELATIVE) 	13.215 	7.382 	1.79
norm2_mask::Size_MatType_NormType::(640x480, 16SC1, NORM_L1) 	8.094 	8.104 	1.00
norm2_mask::Size_MatType_NormType::(640x480, 16SC1, NORM_L1|NORM_RELATIVE) 	14.857 	8.482 	1.75
norm2_mask::Size_MatType_NormType::(640x480, 16SC1, NORM_L2) 	6.814 	5.721 	1.19
norm2_mask::Size_MatType_NormType::(640x480, 16SC1, NORM_L2|NORM_RELATIVE) 	9.322 	6.274 	1.49
norm2_mask::Size_MatType_NormType::(640x480, 32SC1, NORM_INF) 	7.203 	7.203 	1.00
norm2_mask::Size_MatType_NormType::(640x480, 32SC1, NORM_INF|NORM_RELATIVE) 	13.017 	7.686 	1.69
norm2_mask::Size_MatType_NormType::(640x480, 32SC1, NORM_L1) 	6.994 	6.126 	1.14
norm2_mask::Size_MatType_NormType::(640x480, 32SC1, NORM_L1|NORM_RELATIVE) 	12.831 	6.755 	1.90
norm2_mask::Size_MatType_NormType::(640x480, 32SC1, NORM_L2) 	6.829 	5.948 	1.15
norm2_mask::Size_MatType_NormType::(640x480, 32SC1, NORM_L2|NORM_RELATIVE) 	12.483 	6.572 	1.90
norm2_mask::Size_MatType_NormType::(640x480, 32FC1, NORM_INF) 	0.646 	0.547 	1.18
norm2_mask::Size_MatType_NormType::(640x480, 32FC1, NORM_INF|NORM_RELATIVE) 	0.870 	0.769 	1.13
norm2_mask::Size_MatType_NormType::(640x480, 32FC1, NORM_L1) 	0.622 	0.494 	1.26
norm2_mask::Size_MatType_NormType::(640x480, 32FC1, NORM_L1|NORM_RELATIVE) 	0.933 	0.795 	1.17
norm2_mask::Size_MatType_NormType::(640x480, 32FC1, NORM_L2) 	0.622 	0.466 	1.33
norm2_mask::Size_MatType_NormType::(640x480, 32FC1, NORM_L2|NORM_RELATIVE) 	0.907 	0.734 	1.23
norm2_mask::Size_MatType_NormType::(640x480, 64FC1, NORM_INF) 	4.909 	4.681 	1.05
norm2_mask::Size_MatType_NormType::(640x480, 64FC1, NORM_INF|NORM_RELATIVE) 	8.663 	5.239 	1.65
norm2_mask::Size_MatType_NormType::(640x480, 64FC1, NORM_L1) 	6.082 	5.300 	1.15
norm2_mask::Size_MatType_NormType::(640x480, 64FC1, NORM_L1|NORM_RELATIVE) 	8.604 	5.872 	1.47
norm2_mask::Size_MatType_NormType::(640x480, 64FC1, NORM_L2) 	6.295 	5.499 	1.14
norm2_mask::Size_MatType_NormType::(640x480, 64FC1, NORM_L2|NORM_RELATIVE) 	8.612 	6.128 	1.41
norm2_mask::Size_MatType_NormType::(640x480, 8UC4, NORM_INF) 	0.643 	0.577 	1.12
norm2_mask::Size_MatType_NormType::(640x480, 8UC4, NORM_INF|NORM_RELATIVE) 	0.878 	0.808 	1.09
norm2_mask::Size_MatType_NormType::(640x480, 8UC4, NORM_L1) 	0.869 	0.805 	1.08
norm2_mask::Size_MatType_NormType::(640x480, 8UC4, NORM_L1|NORM_RELATIVE) 	1.502 	1.108 	1.36
norm2_mask::Size_MatType_NormType::(640x480, 8UC4, NORM_L2) 	1.001 	0.948 	1.06
norm2_mask::Size_MatType_NormType::(640x480, 8UC4, NORM_L2|NORM_RELATIVE) 	1.802 	1.445 	1.25
norm2_mask::Size_MatType_NormType::(1280x720, 8UC1, NORM_INF) 	0.510 	0.467 	1.09
norm2_mask::Size_MatType_NormType::(1280x720, 8UC1, NORM_INF|NORM_RELATIVE) 	0.720 	0.731 	0.99
norm2_mask::Size_MatType_NormType::(1280x720, 8UC1, NORM_L1) 	0.575 	0.582 	0.99
norm2_mask::Size_MatType_NormType::(1280x720, 8UC1, NORM_L1|NORM_RELATIVE) 	0.992 	0.837 	1.19
norm2_mask::Size_MatType_NormType::(1280x720, 8UC1, NORM_L2) 	0.659 	0.664 	0.99
norm2_mask::Size_MatType_NormType::(1280x720, 8UC1, NORM_L2|NORM_RELATIVE) 	1.163 	0.993 	1.17
norm2_mask::Size_MatType_NormType::(1280x720, 8SC1, NORM_INF) 	21.514 	21.520 	1.00
norm2_mask::Size_MatType_NormType::(1280x720, 8SC1, NORM_INF|NORM_RELATIVE) 	39.737 	22.418 	1.77
norm2_mask::Size_MatType_NormType::(1280x720, 8SC1, NORM_L1) 	23.680 	23.704 	1.00
norm2_mask::Size_MatType_NormType::(1280x720, 8SC1, NORM_L1|NORM_RELATIVE) 	44.721 	24.692 	1.81
norm2_mask::Size_MatType_NormType::(1280x720, 8SC1, NORM_L2) 	22.540 	22.573 	1.00
norm2_mask::Size_MatType_NormType::(1280x720, 8SC1, NORM_L2|NORM_RELATIVE) 	40.657 	23.586 	1.72
norm2_mask::Size_MatType_NormType::(1280x720, 16UC1, NORM_INF) 	21.020 	21.037 	1.00
norm2_mask::Size_MatType_NormType::(1280x720, 16UC1, NORM_INF|NORM_RELATIVE) 	36.126 	21.955 	1.65
norm2_mask::Size_MatType_NormType::(1280x720, 16UC1, NORM_L1) 	24.288 	24.306 	1.00
norm2_mask::Size_MatType_NormType::(1280x720, 16UC1, NORM_L1|NORM_RELATIVE) 	41.622 	25.382 	1.64
norm2_mask::Size_MatType_NormType::(1280x720, 16UC1, NORM_L2) 	20.425 	17.198 	1.19
norm2_mask::Size_MatType_NormType::(1280x720, 16UC1, NORM_L2|NORM_RELATIVE) 	39.087 	18.785 	2.08
norm2_mask::Size_MatType_NormType::(1280x720, 16SC1, NORM_INF) 	21.013 	21.000 	1.00
norm2_mask::Size_MatType_NormType::(1280x720, 16SC1, NORM_INF|NORM_RELATIVE) 	39.581 	22.078 	1.79
norm2_mask::Size_MatType_NormType::(1280x720, 16SC1, NORM_L1) 	24.283 	24.333 	1.00
norm2_mask::Size_MatType_NormType::(1280x720, 16SC1, NORM_L1|NORM_RELATIVE) 	44.556 	25.432 	1.75
norm2_mask::Size_MatType_NormType::(1280x720, 16SC1, NORM_L2) 	20.435 	17.174 	1.19
norm2_mask::Size_MatType_NormType::(1280x720, 16SC1, NORM_L2|NORM_RELATIVE) 	27.900 	18.773 	1.49
norm2_mask::Size_MatType_NormType::(1280x720, 32SC1, NORM_INF) 	21.602 	21.611 	1.00
norm2_mask::Size_MatType_NormType::(1280x720, 32SC1, NORM_INF|NORM_RELATIVE) 	38.980 	23.013 	1.69
norm2_mask::Size_MatType_NormType::(1280x720, 32SC1, NORM_L1) 	20.926 	18.365 	1.14
norm2_mask::Size_MatType_NormType::(1280x720, 32SC1, NORM_L1|NORM_RELATIVE) 	38.464 	20.171 	1.91
norm2_mask::Size_MatType_NormType::(1280x720, 32SC1, NORM_L2) 	20.427 	17.860 	1.14
norm2_mask::Size_MatType_NormType::(1280x720, 32SC1, NORM_L2|NORM_RELATIVE) 	37.362 	19.666 	1.90
norm2_mask::Size_MatType_NormType::(1280x720, 32FC1, NORM_INF) 	1.605 	1.656 	0.97
norm2_mask::Size_MatType_NormType::(1280x720, 32FC1, NORM_INF|NORM_RELATIVE) 	2.301 	2.374 	0.97
norm2_mask::Size_MatType_NormType::(1280x720, 32FC1, NORM_L1) 	1.510 	1.479 	1.02
norm2_mask::Size_MatType_NormType::(1280x720, 32FC1, NORM_L1|NORM_RELATIVE) 	2.486 	2.465 	1.01
norm2_mask::Size_MatType_NormType::(1280x720, 32FC1, NORM_L2) 	1.462 	1.501 	0.97
norm2_mask::Size_MatType_NormType::(1280x720, 32FC1, NORM_L2|NORM_RELATIVE) 	2.301 	2.309 	1.00
norm2_mask::Size_MatType_NormType::(1280x720, 64FC1, NORM_INF) 	14.681 	14.000 	1.05
norm2_mask::Size_MatType_NormType::(1280x720, 64FC1, NORM_INF|NORM_RELATIVE) 	25.911 	15.634 	1.66
norm2_mask::Size_MatType_NormType::(1280x720, 64FC1, NORM_L1) 	18.212 	15.894 	1.15
norm2_mask::Size_MatType_NormType::(1280x720, 64FC1, NORM_L1|NORM_RELATIVE) 	25.735 	17.489 	1.47
norm2_mask::Size_MatType_NormType::(1280x720, 64FC1, NORM_L2) 	18.809 	16.477 	1.14
norm2_mask::Size_MatType_NormType::(1280x720, 64FC1, NORM_L2|NORM_RELATIVE) 	25.857 	18.246 	1.42
norm2_mask::Size_MatType_NormType::(1280x720, 8UC4, NORM_INF) 	1.796 	1.734 	1.04
norm2_mask::Size_MatType_NormType::(1280x720, 8UC4, NORM_INF|NORM_RELATIVE) 	2.487 	2.451 	1.01
norm2_mask::Size_MatType_NormType::(1280x720, 8UC4, NORM_L1) 	2.600 	2.421 	1.07
norm2_mask::Size_MatType_NormType::(1280x720, 8UC4, NORM_L1|NORM_RELATIVE) 	4.469 	3.348 	1.33
norm2_mask::Size_MatType_NormType::(1280x720, 8UC4, NORM_L2) 	2.997 	2.851 	1.05
norm2_mask::Size_MatType_NormType::(1280x720, 8UC4, NORM_L2|NORM_RELATIVE) 	5.398 	4.331 	1.25
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC1, NORM_INF) 	1.111 	1.096 	1.01
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC1, NORM_INF|NORM_RELATIVE) 	1.704 	1.704 	1.00
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC1, NORM_L1) 	1.341 	1.318 	1.02
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC1, NORM_L1|NORM_RELATIVE) 	2.265 	1.944 	1.16
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC1, NORM_L2) 	1.522 	1.496 	1.02
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC1, NORM_L2|NORM_RELATIVE) 	2.638 	2.270 	1.16
norm2_mask::Size_MatType_NormType::(1920x1080, 8SC1, NORM_INF) 	48.440 	48.384 	1.00
norm2_mask::Size_MatType_NormType::(1920x1080, 8SC1, NORM_INF|NORM_RELATIVE) 	89.501 	50.480 	1.77
norm2_mask::Size_MatType_NormType::(1920x1080, 8SC1, NORM_L1) 	53.285 	53.284 	1.00
norm2_mask::Size_MatType_NormType::(1920x1080, 8SC1, NORM_L1|NORM_RELATIVE) 	100.549 	55.397 	1.82
norm2_mask::Size_MatType_NormType::(1920x1080, 8SC1, NORM_L2) 	50.694 	50.715 	1.00
norm2_mask::Size_MatType_NormType::(1920x1080, 8SC1, NORM_L2|NORM_RELATIVE) 	91.528 	52.998 	1.73
norm2_mask::Size_MatType_NormType::(1920x1080, 16UC1, NORM_INF) 	47.299 	47.223 	1.00
norm2_mask::Size_MatType_NormType::(1920x1080, 16UC1, NORM_INF|NORM_RELATIVE) 	81.259 	49.382 	1.65
norm2_mask::Size_MatType_NormType::(1920x1080, 16UC1, NORM_L1) 	54.660 	54.675 	1.00
norm2_mask::Size_MatType_NormType::(1920x1080, 16UC1, NORM_L1|NORM_RELATIVE) 	93.558 	57.010 	1.64
norm2_mask::Size_MatType_NormType::(1920x1080, 16UC1, NORM_L2) 	46.021 	38.645 	1.19
norm2_mask::Size_MatType_NormType::(1920x1080, 16UC1, NORM_L2|NORM_RELATIVE) 	87.919 	42.216 	2.08
norm2_mask::Size_MatType_NormType::(1920x1080, 16SC1, NORM_INF) 	47.296 	47.237 	1.00
norm2_mask::Size_MatType_NormType::(1920x1080, 16SC1, NORM_INF|NORM_RELATIVE) 	89.059 	49.667 	1.79
norm2_mask::Size_MatType_NormType::(1920x1080, 16SC1, NORM_L1) 	54.653 	54.703 	1.00
norm2_mask::Size_MatType_NormType::(1920x1080, 16SC1, NORM_L1|NORM_RELATIVE) 	100.194 	57.131 	1.75
norm2_mask::Size_MatType_NormType::(1920x1080, 16SC1, NORM_L2) 	46.041 	38.634 	1.19
norm2_mask::Size_MatType_NormType::(1920x1080, 16SC1, NORM_L2|NORM_RELATIVE) 	62.815 	42.247 	1.49
norm2_mask::Size_MatType_NormType::(1920x1080, 32SC1, NORM_INF) 	48.454 	48.512 	1.00
norm2_mask::Size_MatType_NormType::(1920x1080, 32SC1, NORM_INF|NORM_RELATIVE) 	87.543 	51.687 	1.69
norm2_mask::Size_MatType_NormType::(1920x1080, 32SC1, NORM_L1) 	47.068 	41.233 	1.14
norm2_mask::Size_MatType_NormType::(1920x1080, 32SC1, NORM_L1|NORM_RELATIVE) 	86.458 	45.337 	1.91
norm2_mask::Size_MatType_NormType::(1920x1080, 32SC1, NORM_L2) 	45.875 	40.183 	1.14
norm2_mask::Size_MatType_NormType::(1920x1080, 32SC1, NORM_L2|NORM_RELATIVE) 	83.970 	44.243 	1.90
norm2_mask::Size_MatType_NormType::(1920x1080, 32FC1, NORM_INF) 	3.642 	3.831 	0.95
norm2_mask::Size_MatType_NormType::(1920x1080, 32FC1, NORM_INF|NORM_RELATIVE) 	5.212 	5.414 	0.96
norm2_mask::Size_MatType_NormType::(1920x1080, 32FC1, NORM_L1) 	3.369 	3.317 	1.02
norm2_mask::Size_MatType_NormType::(1920x1080, 32FC1, NORM_L1|NORM_RELATIVE) 	5.513 	5.565 	0.99
norm2_mask::Size_MatType_NormType::(1920x1080, 32FC1, NORM_L2) 	3.257 	3.387 	0.96
norm2_mask::Size_MatType_NormType::(1920x1080, 32FC1, NORM_L2|NORM_RELATIVE) 	5.123 	5.348 	0.96
norm2_mask::Size_MatType_NormType::(1920x1080, 64FC1, NORM_INF) 	33.009 	31.492 	1.05
norm2_mask::Size_MatType_NormType::(1920x1080, 64FC1, NORM_INF|NORM_RELATIVE) 	58.249 	35.050 	1.66
norm2_mask::Size_MatType_NormType::(1920x1080, 64FC1, NORM_L1) 	41.048 	35.677 	1.15
norm2_mask::Size_MatType_NormType::(1920x1080, 64FC1, NORM_L1|NORM_RELATIVE) 	57.869 	39.275 	1.47
norm2_mask::Size_MatType_NormType::(1920x1080, 64FC1, NORM_L2) 	42.356 	37.063 	1.14
norm2_mask::Size_MatType_NormType::(1920x1080, 64FC1, NORM_L2|NORM_RELATIVE) 	58.261 	40.990 	1.42
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC4, NORM_INF) 	3.783 	3.726 	1.02
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC4, NORM_INF|NORM_RELATIVE) 	5.472 	5.388 	1.02
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC4, NORM_L1) 	5.777 	5.415 	1.07
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC4, NORM_L1|NORM_RELATIVE) 	9.980 	7.457 	1.34
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC4, NORM_L2) 	6.698 	6.404 	1.05
norm2_mask::Size_MatType_NormType::(1920x1080, 8UC4, NORM_L2|NORM_RELATIVE) 	12.074 	9.688 	1.25

fengyuentau · 2025-03-18T06:38:07Z

The patch birngs speedup for many cases, but there are some strange regressions against current 4.x:

@asmorkalov Your results are about norm2. Norm2 should not be affected by this patch. I will also improve norm2 after this patch is merged.

core: improve norm of hal rvv opencv#26991 Merge with opencv/opencv_extra#1241 See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [x] I agree to contribute to the project under Apache 2 License. - [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [x] The PR is proposed to the proper branch - [ ] There is a reference to the original bug report and related work - [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [ ] The feature is well documented and sample code can be built with the project CMake

5.x merge 4.x: merge changes of norm and norm_diff in hal rvv from 4.x #27261 Merge with opencv/opencv_extra#1251 No related changes in contrib #26991 from fengyuentau:4x/core/norm2hal_rvv #27045 from fengyuentau:4x/hal_rvv/normDiff Previous "Merge 4.x" on norm_diff vectorization: #27068

fengyuentau added optimization category: core platform: riscv labels Feb 28, 2025

fengyuentau added this to the 4.12.0 milestone Feb 28, 2025

fengyuentau requested review from asmorkalov and vpisarev February 28, 2025 10:36

vpisarev reviewed Mar 5, 2025

View reviewed changes

modules/core/src/hal_replacement.hpp Outdated Show resolved Hide resolved

This comment was marked as outdated.

Sign in to view

fengyuentau force-pushed the 4x/core/norm2hal_rvv branch from 99b991f to 23a8915 Compare March 8, 2025 16:21

This comment was marked as outdated.

Sign in to view

asmorkalov reviewed Mar 10, 2025

View reviewed changes

opencv-alalek reviewed Mar 14, 2025

View reviewed changes

fengyuentau changed the title ~~core: introduced cv_hal_getNormFunc as a better solution and added all implementations for norm~~ core: improve norm of hal rvv Mar 14, 2025

fengyuentau added 9 commits March 17, 2025 14:31

introduced cv_hal_getNormFunc to replace cv_hal_norm; added complete …

8a3e9f6

…implementation with rvv covering different data types, channels and norm types

rename norm function pointer

86b2a97

drop norm.rvv1p0.hpp completely

8567ed2

update perf data

abd0954

fix perf regressions of norm l1 and l2 kernels of 8u, 8s

5061f49

boost norm l1 kernels (ushort, short)

1440b57

further optimized masked norm L1 (8u, 8s, 16u, 16s)

81a17dc

further optimized masked norm L2 (8u, 8s)

c0ba019

revert to the old norm hal interface; implemented simplified iterator

dc31c5d

fengyuentau force-pushed the 4x/core/norm2hal_rvv branch from ef8787e to dc31c5d Compare March 17, 2025 06:34

fengyuentau requested a review from asmorkalov March 18, 2025 06:04

asmorkalov approved these changes Mar 18, 2025

View reviewed changes

asmorkalov merged commit 8207549 into opencv:4.x Mar 18, 2025
43 of 55 checks passed

fengyuentau deleted the 4x/core/norm2hal_rvv branch March 18, 2025 07:37

fengyuentau mentioned this pull request Apr 27, 2025

5.x merge 4.x: merge changes of norm and norm_diff in hal rvv from 4.x #27261

Merged

asmorkalov mentioned this pull request Apr 29, 2025

5.x merge 4.x #27265

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

core: improve norm of hal rvv#26991

core: improve norm of hal rvv#26991
asmorkalov merged 9 commits intoopencv:4.xfrom
fengyuentau:4x/core/norm2hal_rvv

fengyuentau commented Feb 28, 2025 •

edited

Loading

Uh oh!

Uh oh!

This comment was marked as outdated.

This comment was marked as outdated.

fengyuentau commented Mar 8, 2025

Uh oh!

This comment was marked as outdated.

fengyuentau commented Mar 9, 2025

Uh oh!

asmorkalov Mar 10, 2025

Uh oh!

opencv-alalek Mar 14, 2025

Uh oh!

fengyuentau Mar 14, 2025

Uh oh!

fengyuentau commented Mar 18, 2025

Uh oh!

asmorkalov commented Mar 18, 2025

Uh oh!

fengyuentau commented Mar 18, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

		inline int hal_ni_norm(const uchar* src, size_t src_step, const uchar* mask, size_t mask_step, int width,
		int height, int type, int norm_type, double* result) { return CV_HAL_ERROR_NOT_IMPLEMENTED; }

Uh oh!

Conversation

fengyuentau commented Feb 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Pull Request Readiness Checklist

Uh oh!

Uh oh!

This comment was marked as outdated.

This comment was marked as outdated.

fengyuentau commented Mar 8, 2025

Uh oh!

This comment was marked as outdated.

fengyuentau commented Mar 9, 2025

Uh oh!

asmorkalov Mar 10, 2025

Choose a reason for hiding this comment

Uh oh!

opencv-alalek Mar 14, 2025

Choose a reason for hiding this comment

Uh oh!

fengyuentau Mar 14, 2025

Choose a reason for hiding this comment

Uh oh!

fengyuentau commented Mar 18, 2025

Uh oh!

asmorkalov commented Mar 18, 2025

Uh oh!

fengyuentau commented Mar 18, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

fengyuentau commented Feb 28, 2025 •

edited

Loading